dry run

.github/workflows/deployment.yml

@@ -29,27 +29,15 @@ jobs:
       build-backend-craft: ${{ steps.check.outputs.build-backend-craft }}
       build-model-server: ${{ steps.check.outputs.build-model-server }}
       is-cloud-tag: ${{ steps.check.outputs.is-cloud-tag }}
+      is-stable: ${{ steps.check.outputs.is-stable }}
       is-beta: ${{ steps.check.outputs.is-beta }}
+      is-stable-standalone: ${{ steps.check.outputs.is-stable-standalone }}
       is-beta-standalone: ${{ steps.check.outputs.is-beta-standalone }}
       is-craft-latest: ${{ steps.check.outputs.is-craft-latest }}
-      is-latest: ${{ steps.check.outputs.is-latest }}
       is-test-run: ${{ steps.check.outputs.is-test-run }}
       sanitized-tag: ${{ steps.check.outputs.sanitized-tag }}
       short-sha: ${{ steps.check.outputs.short-sha }}
     steps:
-      - name: Checkout (for git tags)
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # ratchet:actions/checkout@v6
-        with:
-          persist-credentials: false
-          fetch-depth: 0
-          fetch-tags: true
-
-      - name: Setup uv
-        uses: astral-sh/setup-uv@5a095e7a2014a4212f075830d4f7277575a9d098 # ratchet:astral-sh/setup-uv@v7
-        with:
-          version: "0.9.9"
-          enable-cache: false
-
       - name: Check which components to build and version info
         id: check
         env:
@@ -66,9 +54,9 @@ jobs:
           IS_VERSION_TAG=false
           IS_STABLE=false
           IS_BETA=false
+          IS_STABLE_STANDALONE=false
           IS_BETA_STANDALONE=false
           IS_CRAFT_LATEST=false
-          IS_LATEST=false
           IS_PROD_TAG=false
           IS_TEST_RUN=false
           BUILD_DESKTOP=false
@@ -116,22 +104,13 @@ jobs:
           fi
 
           # Standalone version checks (for backend/model-server - version excluding cloud tags)
+          if [[ "$IS_STABLE" == "true" ]] && [[ "$IS_CLOUD" != "true" ]]; then
+            IS_STABLE_STANDALONE=true
+          fi
           if [[ "$IS_BETA" == "true" ]] && [[ "$IS_CLOUD" != "true" ]]; then
             IS_BETA_STANDALONE=true
           fi
 
-          # Determine if this tag should get the "latest" Docker tag.
-          # Only the highest semver stable tag (vX.Y.Z exactly) gets "latest".
-          if [[ "$IS_STABLE" == "true" ]]; then
-            HIGHEST_STABLE=$(uv run --no-sync --with onyx-devtools ods latest-stable-tag) || {
-              echo "::error::Failed to determine highest stable tag via 'ods latest-stable-tag'"
-              exit 1
-            }
-            if [[ "$TAG" == "$HIGHEST_STABLE" ]]; then
-              IS_LATEST=true
-            fi
-          fi
-
           # Determine if this is a production tag
           # Production tags are: version tags (v1.2.3*) or nightly tags
           if [[ "$IS_VERSION_TAG" == "true" ]] || [[ "$IS_NIGHTLY" == "true" ]]; then
@@ -150,10 +129,11 @@ jobs:
             echo "build-backend-craft=$BUILD_BACKEND_CRAFT"
             echo "build-model-server=$BUILD_MODEL_SERVER"
             echo "is-cloud-tag=$IS_CLOUD"
+            echo "is-stable=$IS_STABLE"
             echo "is-beta=$IS_BETA"
+            echo "is-stable-standalone=$IS_STABLE_STANDALONE"
             echo "is-beta-standalone=$IS_BETA_STANDALONE"
             echo "is-craft-latest=$IS_CRAFT_LATEST"
-            echo "is-latest=$IS_LATEST"
             echo "is-test-run=$IS_TEST_RUN"
             echo "sanitized-tag=$SANITIZED_TAG"
             echo "short-sha=$SHORT_SHA"
@@ -620,7 +600,7 @@ jobs:
             latest=false
           tags: |
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run == 'true' && format('web-{0}', needs.determine-builds.outputs.sanitized-tag) || github.ref_name }}
-            type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-latest == 'true' && 'latest' || '' }}
+            type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-stable == 'true' && 'latest' || '' }}
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && env.EDGE_TAG == 'true' && 'edge' || '' }}
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-beta == 'true' && 'beta' || '' }}
 
@@ -1057,7 +1037,7 @@ jobs:
             latest=false
           tags: |
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run == 'true' && format('backend-{0}', needs.determine-builds.outputs.sanitized-tag) || github.ref_name }}
-            type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-latest == 'true' && 'latest' || '' }}
+            type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-stable-standalone == 'true' && 'latest' || '' }}
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && env.EDGE_TAG == 'true' && 'edge' || '' }}
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-beta-standalone == 'true' && 'beta' || '' }}
 
@@ -1493,7 +1473,7 @@ jobs:
             latest=false
           tags: |
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run == 'true' && format('model-server-{0}', needs.determine-builds.outputs.sanitized-tag) || github.ref_name }}
-            type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-latest == 'true' && 'latest' || '' }}
+            type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-stable-standalone == 'true' && 'latest' || '' }}
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && env.EDGE_TAG == 'true' && 'edge' || '' }}
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-beta-standalone == 'true' && 'beta' || '' }}
 

.github/workflows/storybook-deploy.yml

@@ -48,7 +48,7 @@ jobs:
 
       - name: Deploy to Vercel (Production)
         working-directory: web
-        run: npx --yes "$VERCEL_CLI" deploy storybook-static/ --prod --yes --token="$VERCEL_TOKEN"
+        run: npx --yes "$VERCEL_CLI" deploy storybook-static/ --prod --yes
 
   notify-slack-on-failure:
     needs: Deploy-Storybook

backend/alembic/versions/27fb147a843f_add_timestamps_to_user_table.py

@@ -1,43 +0,0 @@
-"""add timestamps to user table
-
-Revision ID: 27fb147a843f
-Revises: b5c4d7e8f9a1
-Create Date: 2026-03-08 17:18:40.828644
-
-"""
-
-from alembic import op
-import sqlalchemy as sa
-
-
-# revision identifiers, used by Alembic.
-revision = "27fb147a843f"
-down_revision = "b5c4d7e8f9a1"
-branch_labels = None
-depends_on = None
-
-
-def upgrade() -> None:
-    op.add_column(
-        "user",
-        sa.Column(
-            "created_at",
-            sa.DateTime(timezone=True),
-            server_default=sa.func.now(),
-            nullable=False,
-        ),
-    )
-    op.add_column(
-        "user",
-        sa.Column(
-            "updated_at",
-            sa.DateTime(timezone=True),
-            server_default=sa.func.now(),
-            nullable=False,
-        ),
-    )
-
-
-def downgrade() -> None:
-    op.drop_column("user", "updated_at")
-    op.drop_column("user", "created_at")

backend/alembic/versions/a3f8b2c1d4e5_add_preferred_response_id_to_chat_message.py

@@ -1,42 +0,0 @@
-"""add multi-model columns to chat_message
-
-Revision ID: a3f8b2c1d4e5
-Revises: 27fb147a843f
-Create Date: 2026-03-12 10:00:00.000000
-
-"""
-
-from alembic import op
-import sqlalchemy as sa
-
-
-# revision identifiers, used by Alembic.
-revision = "a3f8b2c1d4e5"
-down_revision = "27fb147a843f"
-branch_labels = None
-depends_on = None
-
-
-def upgrade() -> None:
-    op.add_column(
-        "chat_message",
-        sa.Column(
-            "preferred_response_id",
-            sa.Integer(),
-            sa.ForeignKey("chat_message.id"),
-            nullable=True,
-        ),
-    )
-    op.add_column(
-        "chat_message",
-        sa.Column(
-            "model_display_name",
-            sa.String(),
-            nullable=True,
-        ),
-    )
-
-
-def downgrade() -> None:
-    op.drop_column("chat_message", "model_display_name")
-    op.drop_column("chat_message", "preferred_response_id")

backend/onyx/chat/models.py

@@ -8,7 +8,6 @@ from onyx.configs.constants import MessageType
 from onyx.context.search.models import SearchDoc
 from onyx.file_store.models import InMemoryChatFile
 from onyx.server.query_and_chat.models import MessageResponseIDInfo
-from onyx.server.query_and_chat.models import MultiModelMessageResponseIDInfo
 from onyx.server.query_and_chat.streaming_models import CitationInfo
 from onyx.server.query_and_chat.streaming_models import GeneratedImage
 from onyx.server.query_and_chat.streaming_models import Packet
@@ -36,13 +35,7 @@ class CreateChatSessionID(BaseModel):
     chat_session_id: UUID
 
 
-AnswerStreamPart = (
-    Packet
-    | MessageResponseIDInfo
-    | MultiModelMessageResponseIDInfo
-    | StreamingError
-    | CreateChatSessionID
-)
+AnswerStreamPart = Packet | MessageResponseIDInfo | StreamingError | CreateChatSessionID
 
 AnswerStream = Iterator[AnswerStreamPart]
 

backend/onyx/connectors/sharepoint/connector.py

@@ -258,10 +258,6 @@ class SharepointConnectorCheckpoint(ConnectorCheckpoint):
     # Track yielded hierarchy nodes by their raw_node_id (URLs) to avoid duplicates
     seen_hierarchy_node_raw_ids: set[str] = Field(default_factory=set)
 
-    # Track yielded document IDs to avoid processing the same document twice.
-    # The Microsoft Graph delta API can return the same item on multiple pages.
-    seen_document_ids: set[str] = Field(default_factory=set)
-
 
 class SharepointAuthMethod(Enum):
     CLIENT_SECRET = "client_secret"
@@ -1561,7 +1557,6 @@ class SharepointConnector(
         checkpoint.current_drive_id = None
         checkpoint.current_drive_web_url = None
         checkpoint.current_drive_delta_next_link = None
-        checkpoint.seen_document_ids.clear()
 
     def _fetch_slim_documents_from_sharepoint(self) -> GenerateSlimDocumentOutput:
         site_descriptors = self.site_descriptors or self.fetch_sites()
@@ -2142,14 +2137,6 @@ class SharepointConnector(
             item_count = 0
             for driveitem in driveitems:
                 item_count += 1
-
-                if driveitem.id and driveitem.id in checkpoint.seen_document_ids:
-                    logger.debug(
-                        f"Skipping duplicate document {driveitem.id} "
-                        f"({driveitem.name})"
-                    )
-                    continue
-
                 driveitem_extension = get_file_ext(driveitem.name)
                 if driveitem_extension not in OnyxFileExtensions.ALL_ALLOWED_EXTENSIONS:
                     logger.warning(
@@ -2202,13 +2189,11 @@ class SharepointConnector(
 
                     if isinstance(doc_or_failure, Document):
                         if doc_or_failure.sections:
-                            checkpoint.seen_document_ids.add(doc_or_failure.id)
                             yield doc_or_failure
                         elif should_yield_if_empty:
                             doc_or_failure.sections = [
                                 TextSection(link=driveitem.web_url, text="")
                             ]
-                            checkpoint.seen_document_ids.add(doc_or_failure.id)
                             yield doc_or_failure
                         else:
                             logger.warning(

backend/onyx/db/llm.py

@@ -25,7 +25,6 @@ from onyx.server.manage.embedding.models import CloudEmbeddingProvider
 from onyx.server.manage.embedding.models import CloudEmbeddingProviderCreationRequest
 from onyx.server.manage.llm.models import LLMProviderUpsertRequest
 from onyx.server.manage.llm.models import LLMProviderView
-from onyx.server.manage.llm.models import SyncModelEntry
 from onyx.utils.logger import setup_logger
 from shared_configs.enums import EmbeddingProvider
 
@@ -370,9 +369,9 @@ def upsert_llm_provider(
 def sync_model_configurations(
     db_session: Session,
     provider_name: str,
-    models: list[SyncModelEntry],
+    models: list[dict],
 ) -> int:
-    """Sync model configurations for a dynamic provider (OpenRouter, Bedrock, Ollama, etc.).
+    """Sync model configurations for a dynamic provider (OpenRouter, Bedrock, Ollama).
 
     This inserts NEW models from the source API without overwriting existing ones.
     User preferences (is_visible, max_input_tokens) are preserved for existing models.
@@ -380,7 +379,7 @@ def sync_model_configurations(
     Args:
         db_session: Database session
         provider_name: Name of the LLM provider
-        models: List of SyncModelEntry objects describing the fetched models
+        models: List of model dicts with keys: name, display_name, max_input_tokens, supports_image_input
 
     Returns:
         Number of new models added
@@ -394,20 +393,21 @@ def sync_model_configurations(
 
     new_count = 0
     for model in models:
-        if model.name not in existing_names:
+        model_name = model["name"]
+        if model_name not in existing_names:
             # Insert new model with is_visible=False (user must explicitly enable)
             supported_flows = [LLMModelFlowType.CHAT]
-            if model.supports_image_input:
+            if model.get("supports_image_input", False):
                 supported_flows.append(LLMModelFlowType.VISION)
 
             insert_new_model_configuration__no_commit(
                 db_session=db_session,
                 llm_provider_id=provider.id,
-                model_name=model.name,
+                model_name=model_name,
                 supported_flows=supported_flows,
                 is_visible=False,
-                max_input_tokens=model.max_input_tokens,
-                display_name=model.display_name,
+                max_input_tokens=model.get("max_input_tokens"),
+                display_name=model.get("display_name"),
             )
             new_count += 1
 

backend/onyx/db/models.py

@@ -163,8 +163,6 @@ class _EncryptedBase(TypeDecorator):
 
 
 class EncryptedString(_EncryptedBase):
-    # Must redeclare cache_ok in this child class since we explicitly redeclare _is_json
-    cache_ok = True
     _is_json: bool = False
 
     def process_bind_param(
@@ -191,7 +189,6 @@ class EncryptedString(_EncryptedBase):
 
 
 class EncryptedJson(_EncryptedBase):
-    cache_ok = True
     _is_json: bool = True
 
     def process_bind_param(
@@ -339,16 +336,6 @@ class User(SQLAlchemyBaseUserTableUUID, Base):
         TIMESTAMPAware(timezone=True), nullable=True
     )
 
-    created_at: Mapped[datetime.datetime] = mapped_column(
-        DateTime(timezone=True), server_default=func.now(), nullable=False
-    )
-    updated_at: Mapped[datetime.datetime] = mapped_column(
-        DateTime(timezone=True),
-        server_default=func.now(),
-        onupdate=func.now(),
-        nullable=False,
-    )
-
     default_model: Mapped[str] = mapped_column(Text, nullable=True)
     # organized in typical structured fashion
     # formatted as `displayName__provider__modelName`
@@ -2622,18 +2609,6 @@ class ChatMessage(Base):
         ForeignKey("chat_message.id"), nullable=True
     )
 
-    # For multi-model turns: the user message points to which assistant response
-    # was selected as the preferred one to continue the conversation with.
-    # Only set on user messages that triggered a multi-model generation.
-    preferred_response_id: Mapped[int | None] = mapped_column(
-        ForeignKey("chat_message.id"), nullable=True
-    )
-
-    # The display name of the model that generated this assistant message
-    # (e.g. "GPT-4", "Claude Opus"). Used on session reload to label
-    # multi-model response panels and <> navigation arrows.
-    model_display_name: Mapped[str | None] = mapped_column(String, nullable=True)
-
     # Only set on summary messages - the ID of the last message included in this summary
     # Used for chat history compression
     last_summarized_message_id: Mapped[int | None] = mapped_column(
@@ -2708,12 +2683,6 @@ class ChatMessage(Base):
         remote_side="ChatMessage.id",
     )
 
-    preferred_response: Mapped["ChatMessage | None"] = relationship(
-        "ChatMessage",
-        foreign_keys=[preferred_response_id],
-        remote_side="ChatMessage.id",
-    )
-
     # Chat messages only need to know their immediate tool call children
     # If there are nested tool calls, they are stored in the tool_call_children relationship.
     tool_calls: Mapped[list["ToolCall"] | None] = relationship(

backend/onyx/db/users.py

@@ -4,7 +4,6 @@ from uuid import UUID
 
 from fastapi import HTTPException
 from fastapi_users.password import PasswordHelper
-from sqlalchemy import case
 from sqlalchemy import func
 from sqlalchemy import select
 from sqlalchemy.exc import IntegrityError
@@ -12,7 +11,6 @@ from sqlalchemy.orm import Session
 from sqlalchemy.sql import expression
 from sqlalchemy.sql.elements import ColumnElement
 from sqlalchemy.sql.elements import KeyedColumnElement
-from sqlalchemy.sql.expression import or_
 
 from onyx.auth.invited_users import remove_user_from_invited_users
 from onyx.auth.schemas import UserRole
@@ -26,7 +24,6 @@ from onyx.db.models import Persona__User
 from onyx.db.models import SamlAccount
 from onyx.db.models import User
 from onyx.db.models import User__UserGroup
-from onyx.db.models import UserGroup
 from onyx.utils.variable_functionality import fetch_ee_implementation_or_noop
 
 
@@ -165,13 +162,7 @@ def _get_accepted_user_where_clause(
         where_clause.append(User.role != UserRole.EXT_PERM_USER)
 
     if email_filter_string is not None:
-        personal_name_col: KeyedColumnElement[Any] = User.__table__.c.personal_name
-        where_clause.append(
-            or_(
-                email_col.ilike(f"%{email_filter_string}%"),
-                personal_name_col.ilike(f"%{email_filter_string}%"),
-            )
-        )
+        where_clause.append(email_col.ilike(f"%{email_filter_string}%"))
 
     if roles_filter:
         where_clause.append(User.role.in_(roles_filter))
@@ -182,21 +173,6 @@ def _get_accepted_user_where_clause(
     return where_clause
 
 
-def get_all_accepted_users(
-    db_session: Session,
-    include_external: bool = False,
-) -> Sequence[User]:
-    """Returns all accepted users without pagination.
-    Uses the same filtering as the paginated endpoint but without
-    search, role, or active filters."""
-    stmt = select(User)
-    where_clause = _get_accepted_user_where_clause(
-        include_external=include_external,
-    )
-    stmt = stmt.where(*where_clause).order_by(User.email)
-    return db_session.scalars(stmt).unique().all()
-
-
 def get_page_of_filtered_users(
     db_session: Session,
     page_size: int,
@@ -242,41 +218,6 @@ def get_total_filtered_users_count(
     return db_session.scalar(total_count_stmt) or 0
 
 
-def get_user_counts_by_role_and_status(
-    db_session: Session,
-) -> dict[str, dict[str, int]]:
-    """Returns user counts grouped by role and by active/inactive status.
-
-    Excludes API key users, anonymous users, and no-auth placeholder users.
-    Uses a single query with conditional aggregation.
-    """
-    base_where = _get_accepted_user_where_clause()
-    role_col = User.__table__.c.role
-    is_active_col = User.__table__.c.is_active
-
-    stmt = (
-        select(
-            role_col,
-            func.count().label("total"),
-            func.sum(case((is_active_col.is_(True), 1), else_=0)).label("active"),
-            func.sum(case((is_active_col.is_(False), 1), else_=0)).label("inactive"),
-        )
-        .where(*base_where)
-        .group_by(role_col)
-    )
-
-    role_counts: dict[str, int] = {}
-    status_counts: dict[str, int] = {"active": 0, "inactive": 0}
-
-    for role_val, total, active, inactive in db_session.execute(stmt).all():
-        key = role_val.value if hasattr(role_val, "value") else str(role_val)
-        role_counts[key] = total
-        status_counts["active"] += active or 0
-        status_counts["inactive"] += inactive or 0
-
-    return {"role_counts": role_counts, "status_counts": status_counts}
-
-
 def get_user_by_email(email: str, db_session: Session) -> User | None:
     user = (
         db_session.query(User)
@@ -353,23 +294,24 @@ def batch_add_ext_perm_user_if_not_exists(
     lower_emails = [email.lower() for email in emails]
     found_users, missing_lower_emails = _get_users_by_emails(db_session, lower_emails)
 
-    # Use savepoints (begin_nested) so that a failed insert only rolls back
-    # that single user, not the entire transaction. A plain rollback() would
-    # discard all previously flushed users in the same transaction.
-    # We also avoid add_all() because SQLAlchemy 2.0's insertmanyvalues
-    # batch path hits a UUID sentinel mismatch with server_default columns.
+    new_users: list[User] = []
     for email in missing_lower_emails:
-        user = _generate_ext_permissioned_user(email=email)
-        savepoint = db_session.begin_nested()
-        try:
-            db_session.add(user)
-            savepoint.commit()
-        except IntegrityError:
-            savepoint.rollback()
-            if not continue_on_error:
-                raise
+        new_users.append(_generate_ext_permissioned_user(email=email))
 
-    db_session.commit()
+    try:
+        db_session.add_all(new_users)
+        db_session.commit()
+    except IntegrityError:
+        db_session.rollback()
+        if not continue_on_error:
+            raise
+        for user in new_users:
+            try:
+                db_session.add(user)
+                db_session.commit()
+            except IntegrityError:
+                db_session.rollback()
+                continue
     # Fetch all users again to ensure we have the most up-to-date list
     all_users, _ = _get_users_by_emails(db_session, lower_emails)
     return all_users
@@ -416,28 +358,3 @@ def delete_user_from_db(
     # NOTE: edge case may exist with race conditions
     # with this `invited user` scheme generally.
     remove_user_from_invited_users(user_to_delete.email)
-
-
-def batch_get_user_groups(
-    db_session: Session,
-    user_ids: list[UUID],
-) -> dict[UUID, list[tuple[int, str]]]:
-    """Fetch group memberships for a batch of users in a single query.
-    Returns a mapping of user_id -> list of (group_id, group_name) tuples."""
-    if not user_ids:
-        return {}
-
-    rows = db_session.execute(
-        select(
-            User__UserGroup.user_id,
-            UserGroup.id,
-            UserGroup.name,
-        )
-        .join(UserGroup, UserGroup.id == User__UserGroup.user_group_id)
-        .where(User__UserGroup.user_id.in_(user_ids))
-    ).all()
-
-    result: dict[UUID, list[tuple[int, str]]] = {uid: [] for uid in user_ids}
-    for user_id, group_id, group_name in rows:
-        result[user_id].append((group_id, group_name))
-    return result

backend/onyx/document_index/opensearch/opensearch_document_index.py

@@ -433,16 +433,12 @@ class OpenSearchOldDocumentIndex(OldDocumentIndex):
             hidden=fields.hidden if fields else None,
             project_ids=(
                 set(user_fields.user_projects)
-                # NOTE: Empty user_projects is semantically different from None
-                # user_projects.
-                if user_fields and user_fields.user_projects is not None
+                if user_fields and user_fields.user_projects
                 else None
             ),
             persona_ids=(
                 set(user_fields.personas)
-                # NOTE: Empty personas is semantically different from None
-                # personas.
-                if user_fields and user_fields.personas is not None
+                if user_fields and user_fields.personas
                 else None
             ),
         )

backend/onyx/document_index/opensearch/search.py

@@ -255,12 +255,8 @@ class DocumentQuery:
                 f"result window ({DEFAULT_OPENSEARCH_MAX_RESULT_WINDOW})."
             )
 
-        # TODO(andrei, yuhong): We can tune this more dynamically based on
-        # num_hits.
-        max_results_per_subquery = DEFAULT_NUM_HYBRID_SEARCH_CANDIDATES
-
         hybrid_search_subqueries = DocumentQuery._get_hybrid_search_subqueries(
-            query_text, query_vector, vector_candidates=max_results_per_subquery
+            query_text, query_vector
         )
         hybrid_search_filters = DocumentQuery._get_search_filters(
             tenant_state=tenant_state,
@@ -295,7 +291,7 @@ class DocumentQuery:
                 # Sources:
                 # https://docs.opensearch.org/latest/vector-search/ai-search/hybrid-search/pagination/
                 # https://opensearch.org/blog/navigating-pagination-in-hybrid-queries-with-the-pagination_depth-parameter/
-                "pagination_depth": max_results_per_subquery,
+                "pagination_depth": DEFAULT_NUM_HYBRID_SEARCH_CANDIDATES,
                 # Applied to all the sub-queries independently (this avoids
                 # subqueries having a lot of results thrown out during
                 # aggregation).
@@ -738,13 +734,14 @@ class DocumentQuery:
             # document's metadata list.
             filter_clauses.append(_get_tag_filter(tags))
 
-        # Knowledge scope: explicit knowledge attachments restrict what an
-        # assistant can see. When none are set the assistant searches
-        # everything.
+        # Knowledge scope: explicit knowledge attachments restrict what
+        # an assistant can see.  When none are set the assistant
+        # searches everything.
         #
-        # project_id / persona_id are additive: they make overflowing user files
-        # findable but must NOT trigger the restriction on their own (an agent
-        # with no explicit knowledge should search everything).
+        # project_id / persona_id are additive: they make overflowing
+        # user files findable but must NOT trigger the restriction on
+        # their own (an agent with no explicit knowledge should search
+        # everything).
         has_knowledge_scope = (
             attached_document_ids
             or hierarchy_node_ids
@@ -772,8 +769,9 @@ class DocumentQuery:
                 knowledge_filter["bool"]["should"].append(
                     _get_document_set_filter(document_sets)
                 )
-            # Additive: widen scope to also cover overflowing user files, but
-            # only when an explicit restriction is already in effect.
+            # Additive: widen scope to also cover overflowing user
+            # files, but only when an explicit restriction is already
+            # in effect.
             if project_id is not None:
                 knowledge_filter["bool"]["should"].append(
                     _get_user_project_filter(project_id)

backend/onyx/document_index/vespa/index.py

@@ -690,12 +690,9 @@ class VespaIndex(DocumentIndex):
             )
 
         project_ids: set[int] | None = None
-        # NOTE: Empty user_projects is semantically different from None
-        # user_projects.
         if user_fields is not None and user_fields.user_projects is not None:
             project_ids = set(user_fields.user_projects)
         persona_ids: set[int] | None = None
-        # NOTE: Empty personas is semantically different from None personas.
         if user_fields is not None and user_fields.personas is not None:
             persona_ids = set(user_fields.personas)
         update_request = MetadataUpdateRequest(

backend/onyx/indexing/adapters/document_indexing_adapter.py

@@ -123,11 +123,15 @@ class DocumentIndexingBatchAdapter:
         }
 
         doc_id_to_new_chunk_cnt: dict[str, int] = {
-            doc_id: 0 for doc_id in updatable_ids
+            document_id: len(
+                [
+                    chunk
+                    for chunk in chunks_with_embeddings
+                    if chunk.source_document.id == document_id
+                ]
+            )
+            for document_id in updatable_ids
         }
-        for chunk in chunks_with_embeddings:
-            if chunk.source_document.id in doc_id_to_new_chunk_cnt:
-                doc_id_to_new_chunk_cnt[chunk.source_document.id] += 1
 
         # Get ancestor hierarchy node IDs for each document
         doc_id_to_ancestor_ids = self._get_ancestor_ids_for_documents(

backend/onyx/indexing/embedder.py

@@ -16,7 +16,6 @@ from onyx.indexing.models import DocAwareChunk
 from onyx.indexing.models import IndexChunk
 from onyx.natural_language_processing.search_nlp_models import EmbeddingModel
 from onyx.utils.logger import setup_logger
-from onyx.utils.pydantic_util import shallow_model_dump
 from onyx.utils.timing import log_function_time
 from shared_configs.configs import INDEXING_MODEL_SERVER_HOST
 from shared_configs.configs import INDEXING_MODEL_SERVER_PORT
@@ -211,8 +210,8 @@ class DefaultIndexingEmbedder(IndexingEmbedder):
                     )[0]
                     title_embed_dict[title] = title_embedding
 
-            new_embedded_chunk = IndexChunk.model_construct(
-                **shallow_model_dump(chunk),
+            new_embedded_chunk = IndexChunk(
+                **chunk.model_dump(),
                 embeddings=ChunkEmbedding(
                     full_embedding=chunk_embeddings[0],
                     mini_chunk_embeddings=chunk_embeddings[1:],

backend/onyx/indexing/models.py

@@ -12,7 +12,6 @@ from onyx.connectors.models import Document
 from onyx.db.enums import EmbeddingPrecision
 from onyx.db.enums import SwitchoverType
 from onyx.utils.logger import setup_logger
-from onyx.utils.pydantic_util import shallow_model_dump
 from shared_configs.enums import EmbeddingProvider
 from shared_configs.model_server_models import Embedding
 
@@ -134,8 +133,9 @@ class DocMetadataAwareIndexChunk(IndexChunk):
         tenant_id: str,
         ancestor_hierarchy_node_ids: list[int] | None = None,
     ) -> "DocMetadataAwareIndexChunk":
-        return cls.model_construct(
-            **shallow_model_dump(index_chunk),
+        index_chunk_data = index_chunk.model_dump()
+        return cls(
+            **index_chunk_data,
             access=access,
             document_sets=document_sets,
             user_project=user_project,

backend/onyx/llm/constants.py

@@ -43,7 +43,6 @@ WELL_KNOWN_PROVIDER_NAMES = [
     LlmProviderNames.AZURE,
     LlmProviderNames.OLLAMA_CHAT,
     LlmProviderNames.LM_STUDIO,
-    LlmProviderNames.LITELLM_PROXY,
 ]
 
 
@@ -60,7 +59,6 @@ PROVIDER_DISPLAY_NAMES: dict[str, str] = {
     "ollama": "Ollama",
     LlmProviderNames.OLLAMA_CHAT: "Ollama",
     LlmProviderNames.LM_STUDIO: "LM Studio",
-    LlmProviderNames.LITELLM_PROXY: "LiteLLM Proxy",
     "groq": "Groq",
     "anyscale": "Anyscale",
     "deepseek": "DeepSeek",
@@ -111,7 +109,6 @@ AGGREGATOR_PROVIDERS: set[str] = {
     LlmProviderNames.LM_STUDIO,
     LlmProviderNames.VERTEX_AI,
     LlmProviderNames.AZURE,
-    LlmProviderNames.LITELLM_PROXY,
 }
 
 # Model family name mappings for display name generation

backend/onyx/llm/model_metadata_enrichments.json

@@ -3782,6 +3782,16 @@
     "display_name": "Claude Sonnet 3.5",
     "model_vendor": "anthropic"
   },
+  "vertex_ai/claude-3-5-sonnet-v2": {
+    "display_name": "Claude Sonnet 3.5",
+    "model_vendor": "anthropic",
+    "model_version": "v2"
+  },
+  "vertex_ai/claude-3-5-sonnet-v2@20241022": {
+    "display_name": "Claude Sonnet 3.5 v2",
+    "model_vendor": "anthropic",
+    "model_version": "20241022"
+  },
   "vertex_ai/claude-3-5-sonnet@20240620": {
     "display_name": "Claude Sonnet 3.5",
     "model_vendor": "anthropic",

backend/onyx/llm/well_known_providers/constants.py

@@ -11,8 +11,6 @@ OLLAMA_API_KEY_CONFIG_KEY = "OLLAMA_API_KEY"
 LM_STUDIO_PROVIDER_NAME = "lm_studio"
 LM_STUDIO_API_KEY_CONFIG_KEY = "LM_STUDIO_API_KEY"
 
-LITELLM_PROXY_PROVIDER_NAME = "litellm_proxy"
-
 # Providers that use optional Bearer auth from custom_config
 PROVIDERS_WITH_SPECIAL_API_KEY_HANDLING: dict[str, str] = {
     LlmProviderNames.OLLAMA_CHAT: OLLAMA_API_KEY_CONFIG_KEY,

backend/onyx/llm/well_known_providers/llm_provider_options.py

@@ -15,7 +15,6 @@ from onyx.llm.well_known_providers.auto_update_service import (
 from onyx.llm.well_known_providers.constants import ANTHROPIC_PROVIDER_NAME
 from onyx.llm.well_known_providers.constants import AZURE_PROVIDER_NAME
 from onyx.llm.well_known_providers.constants import BEDROCK_PROVIDER_NAME
-from onyx.llm.well_known_providers.constants import LITELLM_PROXY_PROVIDER_NAME
 from onyx.llm.well_known_providers.constants import LM_STUDIO_PROVIDER_NAME
 from onyx.llm.well_known_providers.constants import OLLAMA_PROVIDER_NAME
 from onyx.llm.well_known_providers.constants import OPENAI_PROVIDER_NAME
@@ -48,7 +47,6 @@ def _get_provider_to_models_map() -> dict[str, list[str]]:
         OLLAMA_PROVIDER_NAME: [],  # Dynamic - fetched from Ollama API
         LM_STUDIO_PROVIDER_NAME: [],  # Dynamic - fetched from LM Studio API
         OPENROUTER_PROVIDER_NAME: [],  # Dynamic - fetched from OpenRouter API
-        LITELLM_PROXY_PROVIDER_NAME: [],  # Dynamic - fetched from LiteLLM proxy API
     }
 
 
@@ -333,7 +331,6 @@ def get_provider_display_name(provider_name: str) -> str:
         BEDROCK_PROVIDER_NAME: "Amazon Bedrock",
         VERTEXAI_PROVIDER_NAME: "Google Vertex AI",
         OPENROUTER_PROVIDER_NAME: "OpenRouter",
-        LITELLM_PROXY_PROVIDER_NAME: "LiteLLM Proxy",
     }
 
     if provider_name in _ONYX_PROVIDER_DISPLAY_NAMES:

backend/onyx/onyxbot/slack/constants.py

@@ -1,9 +1,5 @@
-import re
 from enum import Enum
 
-# Matches Slack channel references like <#C097NBWMY8Y> or <#C097NBWMY8Y|channel-name>
-SLACK_CHANNEL_REF_PATTERN = re.compile(r"<#([A-Z0-9]+)(?:\|([^>]+))?>")
-
 LIKE_BLOCK_ACTION_ID = "feedback-like"
 DISLIKE_BLOCK_ACTION_ID = "feedback-dislike"
 SHOW_EVERYONE_ACTION_ID = "show-everyone"

backend/onyx/onyxbot/slack/handlers/handle_regular_answer.py

@@ -18,18 +18,15 @@ from onyx.configs.onyxbot_configs import ONYX_BOT_DISPLAY_ERROR_MSGS
 from onyx.configs.onyxbot_configs import ONYX_BOT_NUM_RETRIES
 from onyx.configs.onyxbot_configs import ONYX_BOT_REACT_EMOJI
 from onyx.context.search.models import BaseFilters
-from onyx.context.search.models import Tag
 from onyx.db.engine.sql_engine import get_session_with_current_tenant
 from onyx.db.models import SlackChannelConfig
 from onyx.db.models import User
 from onyx.db.persona import get_persona_by_id
 from onyx.db.users import get_user_by_email
 from onyx.onyxbot.slack.blocks import build_slack_response_blocks
-from onyx.onyxbot.slack.constants import SLACK_CHANNEL_REF_PATTERN
 from onyx.onyxbot.slack.handlers.utils import send_team_member_message
 from onyx.onyxbot.slack.models import SlackMessageInfo
 from onyx.onyxbot.slack.models import ThreadMessage
-from onyx.onyxbot.slack.utils import get_channel_from_id
 from onyx.onyxbot.slack.utils import get_channel_name_from_id
 from onyx.onyxbot.slack.utils import respond_in_thread_or_channel
 from onyx.onyxbot.slack.utils import SlackRateLimiter
@@ -44,51 +41,6 @@ srl = SlackRateLimiter()
 RT = TypeVar("RT")  # return type
 
 
-def resolve_channel_references(
-    message: str,
-    client: WebClient,
-    logger: OnyxLoggingAdapter,
-) -> tuple[str, list[Tag]]:
-    """Parse Slack channel references from a message, resolve IDs to names,
-    replace the raw markup with readable #channel-name, and return channel tags
-    for search filtering."""
-    tags: list[Tag] = []
-    channel_matches = SLACK_CHANNEL_REF_PATTERN.findall(message)
-    seen_channel_ids: set[str] = set()
-
-    for channel_id, channel_name_from_markup in channel_matches:
-        if channel_id in seen_channel_ids:
-            continue
-        seen_channel_ids.add(channel_id)
-
-        channel_name = channel_name_from_markup or None
-
-        if not channel_name:
-            try:
-                channel_info = get_channel_from_id(client=client, channel_id=channel_id)
-                channel_name = channel_info.get("name") or None
-            except Exception:
-                logger.warning(f"Failed to resolve channel name for ID: {channel_id}")
-
-            if not channel_name:
-                continue
-
-        # Replace raw Slack markup with readable channel name
-        if channel_name_from_markup:
-            message = message.replace(
-                f"<#{channel_id}|{channel_name_from_markup}>",
-                f"#{channel_name}",
-            )
-        else:
-            message = message.replace(
-                f"<#{channel_id}>",
-                f"#{channel_name}",
-            )
-        tags.append(Tag(tag_key="Channel", tag_value=channel_name))
-
-    return message, tags
-
-
 def rate_limits(
     client: WebClient, channel: str, thread_ts: Optional[str]
 ) -> Callable[[Callable[..., RT]], Callable[..., RT]]:
@@ -205,20 +157,6 @@ def handle_regular_answer(
     user_message = messages[-1]
     history_messages = messages[:-1]
 
-    # Resolve any <#CHANNEL_ID> references in the user message to readable
-    # channel names and extract channel tags for search filtering
-    resolved_message, channel_tags = resolve_channel_references(
-        message=user_message.message,
-        client=client,
-        logger=logger,
-    )
-
-    user_message = ThreadMessage(
-        message=resolved_message,
-        sender=user_message.sender,
-        role=user_message.role,
-    )
-
     channel_name, _ = get_channel_name_from_id(
         client=client,
         channel_id=channel,
@@ -269,7 +207,6 @@ def handle_regular_answer(
             source_type=None,
             document_set=document_set_names,
             time_cutoff=None,
-            tags=channel_tags if channel_tags else None,
         )
 
         new_message_request = SendMessageRequest(
@@ -294,16 +231,6 @@ def handle_regular_answer(
             slack_context_str=slack_context_str,
         )
 
-        # If a channel filter was applied but no results were found, override
-        # the LLM response to avoid hallucinated answers about unindexed channels
-        if channel_tags and not answer.citation_info and not answer.top_documents:
-            channel_names = ", ".join(f"#{tag.tag_value}" for tag in channel_tags)
-            answer.answer = (
-                f"No indexed data found for {channel_names}. "
-                "This channel may not be indexed, or there may be no messages "
-                "matching your query within it."
-            )
-
     except Exception as e:
         logger.exception(
             f"Unable to process message - did not successfully answer "
@@ -358,7 +285,6 @@ def handle_regular_answer(
         only_respond_if_citations
         and not answer.citation_info
         and not message_info.bypass_filters
-        and not channel_tags
     ):
         logger.error(
             f"Unable to find citations to answer: '{answer.answer}' - not answering!"

backend/onyx/server/features/build/api/sessions_api.py

@@ -732,7 +732,7 @@ def get_webapp_info(
     return WebappInfo(**webapp_info)
 
 
-@router.get("/{session_id}/webapp-download")
+@router.get("/{session_id}/webapp/download")
 def download_webapp(
     session_id: UUID,
     user: User = Depends(current_user),

backend/onyx/server/features/build/sandbox/kubernetes/docker/templates/outputs/web/package-lock.json

@@ -7424,9 +7424,9 @@
       }
     },
     "node_modules/hono": {
-      "version": "4.12.7",
-      "resolved": "https://registry.npmjs.org/hono/-/hono-4.12.7.tgz",
-      "integrity": "sha512-jq9l1DM0zVIvsm3lv9Nw9nlJnMNPOcAtsbsgiUhWcFzPE99Gvo6yRTlszSLLYacMeQ6quHD6hMfId8crVHvexw==",
+      "version": "4.12.5",
+      "resolved": "https://registry.npmjs.org/hono/-/hono-4.12.5.tgz",
+      "integrity": "sha512-3qq+FUBtlTHhtYxbxheZgY8NIFnkkC/MR8u5TTsr7YZ3wixryQ3cCwn3iZbg8p8B88iDBBAYSfZDS75t8MN7Vg==",
       "license": "MIT",
       "engines": {
         "node": ">=16.9.0"

backend/onyx/server/manage/llm/api.py

@@ -58,9 +58,6 @@ from onyx.llm.well_known_providers.llm_provider_options import (
 from onyx.server.manage.llm.models import BedrockFinalModelResponse
 from onyx.server.manage.llm.models import BedrockModelsRequest
 from onyx.server.manage.llm.models import DefaultModel
-from onyx.server.manage.llm.models import LitellmFinalModelResponse
-from onyx.server.manage.llm.models import LitellmModelDetails
-from onyx.server.manage.llm.models import LitellmModelsRequest
 from onyx.server.manage.llm.models import LLMCost
 from onyx.server.manage.llm.models import LLMProviderDescriptor
 from onyx.server.manage.llm.models import LLMProviderResponse
@@ -75,7 +72,6 @@ from onyx.server.manage.llm.models import OllamaModelsRequest
 from onyx.server.manage.llm.models import OpenRouterFinalModelResponse
 from onyx.server.manage.llm.models import OpenRouterModelDetails
 from onyx.server.manage.llm.models import OpenRouterModelsRequest
-from onyx.server.manage.llm.models import SyncModelEntry
 from onyx.server.manage.llm.models import TestLLMRequest
 from onyx.server.manage.llm.models import VisionProviderResponse
 from onyx.server.manage.llm.utils import generate_bedrock_display_name
@@ -102,34 +98,6 @@ def _mask_string(value: str) -> str:
     return value[:4] + "****" + value[-4:]
 
 
-def _sync_fetched_models(
-    db_session: Session,
-    provider_name: str,
-    models: list[SyncModelEntry],
-    source_label: str,
-) -> None:
-    """Sync fetched models to DB for the given provider.
-
-    Args:
-        db_session: Database session
-        provider_name: Name of the LLM provider
-        models: List of SyncModelEntry objects describing the fetched models
-        source_label: Human-readable label for log messages (e.g. "Bedrock", "LiteLLM")
-    """
-    try:
-        new_count = sync_model_configurations(
-            db_session=db_session,
-            provider_name=provider_name,
-            models=models,
-        )
-        if new_count > 0:
-            logger.info(
-                f"Added {new_count} new {source_label} models to provider '{provider_name}'"
-            )
-    except ValueError as e:
-        logger.warning(f"Failed to sync {source_label} models to DB: {e}")
-
-
 # Keys in custom_config that contain sensitive credentials
 _SENSITIVE_CONFIG_KEYS = {
     "vertex_credentials",
@@ -995,20 +963,27 @@ def get_bedrock_available_models(
 
         # Sync new models to DB if provider_name is specified
         if request.provider_name:
-            _sync_fetched_models(
-                db_session=db_session,
-                provider_name=request.provider_name,
-                models=[
-                    SyncModelEntry(
-                        name=r.name,
-                        display_name=r.display_name,
-                        max_input_tokens=r.max_input_tokens,
-                        supports_image_input=r.supports_image_input,
-                    )
+            try:
+                models_to_sync = [
+                    {
+                        "name": r.name,
+                        "display_name": r.display_name,
+                        "max_input_tokens": r.max_input_tokens,
+                        "supports_image_input": r.supports_image_input,
+                    }
                     for r in results
-                ],
-                source_label="Bedrock",
-            )
+                ]
+                new_count = sync_model_configurations(
+                    db_session=db_session,
+                    provider_name=request.provider_name,
+                    models=models_to_sync,
+                )
+                if new_count > 0:
+                    logger.info(
+                        f"Added {new_count} new Bedrock models to provider '{request.provider_name}'"
+                    )
+            except ValueError as e:
+                logger.warning(f"Failed to sync Bedrock models to DB: {e}")
 
         return results
 
@@ -1126,20 +1101,27 @@ def get_ollama_available_models(
 
     # Sync new models to DB if provider_name is specified
     if request.provider_name:
-        _sync_fetched_models(
-            db_session=db_session,
-            provider_name=request.provider_name,
-            models=[
-                SyncModelEntry(
-                    name=r.name,
-                    display_name=r.display_name,
-                    max_input_tokens=r.max_input_tokens,
-                    supports_image_input=r.supports_image_input,
-                )
+        try:
+            models_to_sync = [
+                {
+                    "name": r.name,
+                    "display_name": r.display_name,
+                    "max_input_tokens": r.max_input_tokens,
+                    "supports_image_input": r.supports_image_input,
+                }
                 for r in sorted_results
-            ],
-            source_label="Ollama",
-        )
+            ]
+            new_count = sync_model_configurations(
+                db_session=db_session,
+                provider_name=request.provider_name,
+                models=models_to_sync,
+            )
+            if new_count > 0:
+                logger.info(
+                    f"Added {new_count} new Ollama models to provider '{request.provider_name}'"
+                )
+        except ValueError as e:
+            logger.warning(f"Failed to sync Ollama models to DB: {e}")
 
     return sorted_results
 
@@ -1228,20 +1210,27 @@ def get_openrouter_available_models(
 
     # Sync new models to DB if provider_name is specified
     if request.provider_name:
-        _sync_fetched_models(
-            db_session=db_session,
-            provider_name=request.provider_name,
-            models=[
-                SyncModelEntry(
-                    name=r.name,
-                    display_name=r.display_name,
-                    max_input_tokens=r.max_input_tokens,
-                    supports_image_input=r.supports_image_input,
-                )
+        try:
+            models_to_sync = [
+                {
+                    "name": r.name,
+                    "display_name": r.display_name,
+                    "max_input_tokens": r.max_input_tokens,
+                    "supports_image_input": r.supports_image_input,
+                }
                 for r in sorted_results
-            ],
-            source_label="OpenRouter",
-        )
+            ]
+            new_count = sync_model_configurations(
+                db_session=db_session,
+                provider_name=request.provider_name,
+                models=models_to_sync,
+            )
+            if new_count > 0:
+                logger.info(
+                    f"Added {new_count} new OpenRouter models to provider '{request.provider_name}'"
+                )
+        except ValueError as e:
+            logger.warning(f"Failed to sync OpenRouter models to DB: {e}")
 
     return sorted_results
 
@@ -1335,119 +1324,26 @@ def get_lm_studio_available_models(
 
     # Sync new models to DB if provider_name is specified
     if request.provider_name:
-        _sync_fetched_models(
-            db_session=db_session,
-            provider_name=request.provider_name,
-            models=[
-                SyncModelEntry(
-                    name=r.name,
-                    display_name=r.display_name,
-                    max_input_tokens=r.max_input_tokens,
-                    supports_image_input=r.supports_image_input,
-                )
-                for r in sorted_results
-            ],
-            source_label="LM Studio",
-        )
-
-    return sorted_results
-
-
-@admin_router.post("/litellm/available-models")
-def get_litellm_available_models(
-    request: LitellmModelsRequest,
-    _: User = Depends(current_admin_user),
-    db_session: Session = Depends(get_session),
-) -> list[LitellmFinalModelResponse]:
-    """Fetch available models from Litellm proxy /v1/models endpoint."""
-    response_json = _get_litellm_models_response(
-        api_key=request.api_key, api_base=request.api_base
-    )
-
-    models = response_json.get("data", [])
-    if not isinstance(models, list) or len(models) == 0:
-        raise OnyxError(
-            OnyxErrorCode.VALIDATION_ERROR,
-            "No models found from your Litellm endpoint",
-        )
-
-    results: list[LitellmFinalModelResponse] = []
-    for model in models:
         try:
-            model_details = LitellmModelDetails.model_validate(model)
-
-            results.append(
-                LitellmFinalModelResponse(
-                    provider_name=model_details.owned_by,
-                    model_name=model_details.id,
-                )
-            )
-        except Exception as e:
-            logger.warning(
-                "Failed to parse Litellm model entry",
-                extra={"error": str(e), "item": str(model)[:1000]},
-            )
-
-    if not results:
-        raise OnyxError(
-            OnyxErrorCode.VALIDATION_ERROR,
-            "No compatible models found from Litellm",
-        )
-
-    sorted_results = sorted(results, key=lambda m: m.model_name.lower())
-
-    # Sync new models to DB if provider_name is specified
-    if request.provider_name:
-        _sync_fetched_models(
-            db_session=db_session,
-            provider_name=request.provider_name,
-            models=[
-                SyncModelEntry(
-                    name=r.model_name,
-                    display_name=r.model_name,
-                )
+            models_to_sync = [
+                {
+                    "name": r.name,
+                    "display_name": r.display_name,
+                    "max_input_tokens": r.max_input_tokens,
+                    "supports_image_input": r.supports_image_input,
+                }
                 for r in sorted_results
-            ],
-            source_label="LiteLLM",
-        )
+            ]
+            new_count = sync_model_configurations(
+                db_session=db_session,
+                provider_name=request.provider_name,
+                models=models_to_sync,
+            )
+            if new_count > 0:
+                logger.info(
+                    f"Added {new_count} new LM Studio models to provider '{request.provider_name}'"
+                )
+        except ValueError as e:
+            logger.warning(f"Failed to sync LM Studio models to DB: {e}")
 
     return sorted_results
-
-
-def _get_litellm_models_response(api_key: str, api_base: str) -> dict:
-    """Perform GET to Litellm proxy /api/v1/models and return parsed JSON."""
-    cleaned_api_base = api_base.strip().rstrip("/")
-    url = f"{cleaned_api_base}/v1/models"
-
-    headers = {
-        "Authorization": f"Bearer {api_key}",
-        "HTTP-Referer": "https://onyx.app",
-        "X-Title": "Onyx",
-    }
-
-    try:
-        response = httpx.get(url, headers=headers, timeout=10.0)
-        response.raise_for_status()
-        return response.json()
-    except httpx.HTTPStatusError as e:
-        if e.response.status_code == 401:
-            raise OnyxError(
-                OnyxErrorCode.VALIDATION_ERROR,
-                "Authentication failed: invalid or missing API key for LiteLLM proxy.",
-            )
-        elif e.response.status_code == 404:
-            raise OnyxError(
-                OnyxErrorCode.VALIDATION_ERROR,
-                f"LiteLLM models endpoint not found at {url}. "
-                "Please verify the API base URL.",
-            )
-        else:
-            raise OnyxError(
-                OnyxErrorCode.BAD_GATEWAY,
-                f"Failed to fetch LiteLLM models: {e}",
-            )
-    except Exception as e:
-        raise OnyxError(
-            OnyxErrorCode.BAD_GATEWAY,
-            f"Failed to fetch LiteLLM models: {e}",
-        )

backend/onyx/server/manage/llm/models.py

@@ -420,32 +420,3 @@ class LLMProviderResponse(BaseModel, Generic[T]):
             default_text=default_text,
             default_vision=default_vision,
         )
-
-
-class SyncModelEntry(BaseModel):
-    """Typed model for syncing fetched models to the DB."""
-
-    name: str
-    display_name: str
-    max_input_tokens: int | None = None
-    supports_image_input: bool = False
-
-
-class LitellmModelsRequest(BaseModel):
-    api_key: str
-    api_base: str
-    provider_name: str | None = None  # Optional: to save models to existing provider
-
-
-class LitellmModelDetails(BaseModel):
-    """Response model for Litellm proxy /api/v1/models endpoint"""
-
-    id: str  # Model ID (e.g. "gpt-4o")
-    object: str  # "model"
-    created: int  # Unix timestamp in seconds
-    owned_by: str  # Provider name (e.g. "openai")
-
-
-class LitellmFinalModelResponse(BaseModel):
-    provider_name: str  # Provider name (e.g. "openai")
-    model_name: str  # Model ID (e.g. "gpt-4o")

backend/onyx/server/manage/users.py

@@ -5,7 +5,6 @@ from datetime import datetime
 from datetime import timedelta
 from datetime import timezone
 from typing import cast
-from uuid import UUID
 
 import jwt
 from email_validator import EmailNotValidError
@@ -19,7 +18,6 @@ from fastapi import Query
 from fastapi import Request
 from fastapi.responses import StreamingResponse
 from pydantic import BaseModel
-from sqlalchemy import select
 from sqlalchemy.orm import Session
 
 from onyx.auth.anonymous_user import fetch_anonymous_user_info
@@ -69,14 +67,11 @@ from onyx.db.user_preferences import update_user_role
 from onyx.db.user_preferences import update_user_shortcut_enabled
 from onyx.db.user_preferences import update_user_temperature_override_enabled
 from onyx.db.user_preferences import update_user_theme_preference
-from onyx.db.users import batch_get_user_groups
 from onyx.db.users import delete_user_from_db
-from onyx.db.users import get_all_accepted_users
 from onyx.db.users import get_all_users
 from onyx.db.users import get_page_of_filtered_users
 from onyx.db.users import get_total_filtered_users_count
 from onyx.db.users import get_user_by_email
-from onyx.db.users import get_user_counts_by_role_and_status
 from onyx.db.users import validate_user_role_update
 from onyx.key_value_store.factory import get_kv_store
 from onyx.redis.redis_pool import get_raw_redis_client
@@ -103,7 +98,6 @@ from onyx.server.manage.models import UserSpecificAssistantPreferences
 from onyx.server.models import FullUserSnapshot
 from onyx.server.models import InvitedUserSnapshot
 from onyx.server.models import MinimalUserSnapshot
-from onyx.server.models import UserGroupInfo
 from onyx.server.usage_limits import is_tenant_on_trial_fn
 from onyx.server.utils import BasicAuthenticationError
 from onyx.utils.logger import setup_logger
@@ -209,91 +203,14 @@ def list_accepted_users(
             total_items=0,
         )
 
-    user_ids = [user.id for user in filtered_accepted_users]
-    groups_by_user = batch_get_user_groups(db_session, user_ids)
-
-    # Batch-fetch SCIM mappings to mark synced users
-    scim_synced_ids: set[UUID] = set()
-    try:
-        from onyx.db.models import ScimUserMapping
-
-        scim_mappings = db_session.scalars(
-            select(ScimUserMapping.user_id).where(ScimUserMapping.user_id.in_(user_ids))
-        ).all()
-        scim_synced_ids = set(scim_mappings)
-    except Exception:
-        logger.warning(
-            "Failed to fetch SCIM mappings; marking all users as non-synced",
-            exc_info=True,
-        )
-
     return PaginatedReturn(
         items=[
-            FullUserSnapshot.from_user_model(
-                user,
-                groups=[
-                    UserGroupInfo(id=gid, name=gname)
-                    for gid, gname in groups_by_user.get(user.id, [])
-                ],
-                is_scim_synced=user.id in scim_synced_ids,
-            )
-            for user in filtered_accepted_users
+            FullUserSnapshot.from_user_model(user) for user in filtered_accepted_users
         ],
         total_items=total_accepted_users_count,
     )
 
 
-@router.get("/manage/users/accepted/all", tags=PUBLIC_API_TAGS)
-def list_all_accepted_users(
-    _: User = Depends(current_admin_user),
-    db_session: Session = Depends(get_session),
-) -> list[FullUserSnapshot]:
-    """Returns all accepted users without pagination.
-    Used by the admin Users page for client-side filtering/sorting."""
-    users = get_all_accepted_users(db_session=db_session)
-
-    if not users:
-        return []
-
-    user_ids = [user.id for user in users]
-    groups_by_user = batch_get_user_groups(db_session, user_ids)
-
-    # Batch-fetch SCIM mappings to mark synced users
-    scim_synced_ids: set[UUID] = set()
-    try:
-        from onyx.db.models import ScimUserMapping
-
-        scim_mappings = db_session.scalars(
-            select(ScimUserMapping.user_id).where(ScimUserMapping.user_id.in_(user_ids))
-        ).all()
-        scim_synced_ids = set(scim_mappings)
-    except Exception:
-        logger.warning(
-            "Failed to fetch SCIM mappings; marking all users as non-synced",
-            exc_info=True,
-        )
-
-    return [
-        FullUserSnapshot.from_user_model(
-            user,
-            groups=[
-                UserGroupInfo(id=gid, name=gname)
-                for gid, gname in groups_by_user.get(user.id, [])
-            ],
-            is_scim_synced=user.id in scim_synced_ids,
-        )
-        for user in users
-    ]
-
-
-@router.get("/manage/users/counts")
-def get_user_counts(
-    _: User = Depends(current_admin_user),
-    db_session: Session = Depends(get_session),
-) -> dict[str, dict[str, int]]:
-    return get_user_counts_by_role_and_status(db_session)
-
-
 @router.get("/manage/users/invited", tags=PUBLIC_API_TAGS)
 def list_invited_users(
     _: User = Depends(current_admin_user),
@@ -352,10 +269,24 @@ def list_all_users(
     if accepted_page is None or invited_page is None or slack_users_page is None:
         return AllUsersResponse(
             accepted=[
-                FullUserSnapshot.from_user_model(user) for user in accepted_users
+                FullUserSnapshot(
+                    id=user.id,
+                    email=user.email,
+                    role=user.role,
+                    is_active=user.is_active,
+                    password_configured=user.password_configured,
+                )
+                for user in accepted_users
             ],
             slack_users=[
-                FullUserSnapshot.from_user_model(user) for user in slack_users
+                FullUserSnapshot(
+                    id=user.id,
+                    email=user.email,
+                    role=user.role,
+                    is_active=user.is_active,
+                    password_configured=user.password_configured,
+                )
+                for user in slack_users
             ],
             invited=[InvitedUserSnapshot(email=email) for email in invited_emails],
             accepted_pages=1,
@@ -365,10 +296,26 @@ def list_all_users(
 
     # Otherwise, return paginated results
     return AllUsersResponse(
-        accepted=[FullUserSnapshot.from_user_model(user) for user in accepted_users][
-            accepted_page * USERS_PAGE_SIZE : (accepted_page + 1) * USERS_PAGE_SIZE
-        ],
-        slack_users=[FullUserSnapshot.from_user_model(user) for user in slack_users][
+        accepted=[
+            FullUserSnapshot(
+                id=user.id,
+                email=user.email,
+                role=user.role,
+                is_active=user.is_active,
+                password_configured=user.password_configured,
+            )
+            for user in accepted_users
+        ][accepted_page * USERS_PAGE_SIZE : (accepted_page + 1) * USERS_PAGE_SIZE],
+        slack_users=[
+            FullUserSnapshot(
+                id=user.id,
+                email=user.email,
+                role=user.role,
+                is_active=user.is_active,
+                password_configured=user.password_configured,
+            )
+            for user in slack_users
+        ][
             slack_users_page
             * USERS_PAGE_SIZE : (slack_users_page + 1)
             * USERS_PAGE_SIZE

backend/onyx/server/models.py

@@ -1,4 +1,3 @@
-import datetime
 from typing import Generic
 from typing import Optional
 from typing import TypeVar
@@ -32,41 +31,21 @@ class MinimalUserSnapshot(BaseModel):
     email: str
 
 
-class UserGroupInfo(BaseModel):
-    id: int
-    name: str
-
-
 class FullUserSnapshot(BaseModel):
     id: UUID
     email: str
     role: UserRole
     is_active: bool
     password_configured: bool
-    personal_name: str | None
-    created_at: datetime.datetime
-    updated_at: datetime.datetime
-    groups: list[UserGroupInfo]
-    is_scim_synced: bool
 
     @classmethod
-    def from_user_model(
-        cls,
-        user: User,
-        groups: list[UserGroupInfo] | None = None,
-        is_scim_synced: bool = False,
-    ) -> "FullUserSnapshot":
+    def from_user_model(cls, user: User) -> "FullUserSnapshot":
         return cls(
             id=user.id,
             email=user.email,
             role=user.role,
             is_active=user.is_active,
             password_configured=user.password_configured,
-            personal_name=user.personal_name,
-            created_at=user.created_at,
-            updated_at=user.updated_at,
-            groups=groups or [],
-            is_scim_synced=is_scim_synced,
         )
 
 

backend/onyx/server/query_and_chat/models.py

@@ -41,16 +41,6 @@ class MessageResponseIDInfo(BaseModel):
     reserved_assistant_message_id: int
 
 
-class MultiModelMessageResponseIDInfo(BaseModel):
-    """Sent at the start of a multi-model streaming response.
-    Contains the user message ID and the reserved assistant message IDs
-    for each model being run in parallel."""
-
-    user_message_id: int | None
-    reserved_assistant_message_ids: list[int]
-    model_names: list[str]
-
-
 class SourceTag(Tag):
     source: DocumentSource
 
@@ -96,10 +86,6 @@ class SendMessageRequest(BaseModel):
     message: str
 
     llm_override: LLMOverride | None = None
-    # For multi-model mode: up to 3 LLM overrides to run in parallel.
-    # When provided with >1 entry, triggers multi-model streaming.
-    # Backward-compat: if only `llm_override` is set, single-model path is used.
-    llm_overrides: list[LLMOverride] | None = None
     # Test-only override for deterministic LiteLLM mock responses.
     mock_llm_response: str | None = None
 
@@ -225,10 +211,6 @@ class ChatMessageDetail(BaseModel):
     error: str | None = None
     current_feedback: str | None = None  # "like" | "dislike" | null
     processing_duration_seconds: float | None = None
-    # For multi-model turns: the preferred assistant response ID (set on user messages only)
-    preferred_response_id: int | None = None
-    # The display name of the model that generated this message (e.g. "GPT-4", "Claude Opus")
-    model_display_name: str | None = None
 
     def model_dump(self, *args: list, **kwargs: dict[str, Any]) -> dict[str, Any]:  # type: ignore
         initial_dict = super().model_dump(mode="json", *args, **kwargs)  # type: ignore

backend/onyx/server/query_and_chat/placement.py

@@ -8,6 +8,3 @@ class Placement(BaseModel):
     tab_index: int = 0
     # Used for tools/agents that call other tools, this currently doesn't support nested agents but can be added later
     sub_turn_index: int | None = None
-    # For multi-model streaming: identifies which model (0, 1, 2) this packet belongs to.
-    # None for single-model (default) responses.
-    model_index: int | None = None

backend/onyx/utils/pydantic_util.py

@@ -1,13 +0,0 @@
-from typing import Any
-
-from pydantic import BaseModel
-
-
-def shallow_model_dump(model_instance: BaseModel) -> dict[str, Any]:
-    """Like model_dump(), but returns references to field values instead of
-    deep copies. Use with model_construct() to avoid unnecessary memory
-    duplication when building subclass instances."""
-    return {
-        field_name: getattr(model_instance, field_name)
-        for field_name in model_instance.__class__.model_fields
-    }

backend/requirements/default.txt

@@ -750,7 +750,7 @@ pypandoc-binary==1.16.2
     # via onyx
 pyparsing==3.2.5
     # via httplib2
-pypdf==6.8.0
+pypdf==6.7.5
     # via
     #   onyx
     #   unstructured-client
@@ -1020,7 +1020,7 @@ toolz==1.1.0
     #   dask
     #   distributed
     #   partd
-tornado==6.5.5
+tornado==6.5.2
     # via distributed
 tqdm==4.67.1
     # via

backend/requirements/dev.txt

@@ -263,7 +263,7 @@ oauthlib==3.2.2
     # via
     #   kubernetes
     #   requests-oauthlib
-onyx-devtools==0.7.0
+onyx-devtools==0.6.3
     # via onyx
 openai==2.14.0
     # via
@@ -406,7 +406,7 @@ referencing==0.36.2
     #   jsonschema-specifications
 regex==2025.11.3
     # via tiktoken
-release-tag==0.5.2
+release-tag==0.4.3
     # via onyx
 reorder-python-imports-black==3.14.0
     # via onyx
@@ -466,7 +466,7 @@ tokenizers==0.21.4
     # via
     #   cohere
     #   litellm
-tornado==6.5.5
+tornado==6.5.2
     # via
     #   ipykernel
     #   jupyter-client

backend/tests/external_dependency_unit/document_index/test_document_index_old.py

@@ -1,398 +0,0 @@
-"""External dependency tests for the old DocumentIndex interface.
-
-These tests assume Vespa and OpenSearch are running.
-
-TODO(ENG-3764)(andrei): Consolidate some of these test fixtures.
-"""
-
-import os
-import time
-import uuid
-from collections.abc import Generator
-from unittest.mock import patch
-
-import httpx
-import pytest
-
-from onyx.access.models import DocumentAccess
-from onyx.configs.constants import DocumentSource
-from onyx.connectors.models import Document
-from onyx.context.search.models import IndexFilters
-from onyx.db.enums import EmbeddingPrecision
-from onyx.document_index.interfaces import DocumentIndex
-from onyx.document_index.interfaces import IndexBatchParams
-from onyx.document_index.interfaces import VespaChunkRequest
-from onyx.document_index.interfaces import VespaDocumentUserFields
-from onyx.document_index.opensearch.client import wait_for_opensearch_with_timeout
-from onyx.document_index.opensearch.opensearch_document_index import (
-    OpenSearchOldDocumentIndex,
-)
-from onyx.document_index.vespa.index import VespaIndex
-from onyx.document_index.vespa.shared_utils.utils import get_vespa_http_client
-from onyx.document_index.vespa.shared_utils.utils import wait_for_vespa_with_timeout
-from onyx.indexing.models import ChunkEmbedding
-from onyx.indexing.models import DocMetadataAwareIndexChunk
-from shared_configs.configs import MULTI_TENANT
-from shared_configs.contextvars import CURRENT_TENANT_ID_CONTEXTVAR
-from shared_configs.contextvars import get_current_tenant_id
-from tests.external_dependency_unit.constants import TEST_TENANT_ID
-
-
-@pytest.fixture(scope="module")
-def opensearch_available() -> Generator[None, None, None]:
-    """Verifies OpenSearch is running, fails the test if not."""
-    if not wait_for_opensearch_with_timeout():
-        pytest.fail("OpenSearch is not available.")
-    yield  # Test runs here.
-
-
-@pytest.fixture(scope="module")
-def test_index_name() -> Generator[str, None, None]:
-    yield f"test_index_{uuid.uuid4().hex[:8]}"  # Test runs here.
-
-
-@pytest.fixture(scope="module")
-def tenant_context() -> Generator[None, None, None]:
-    """Sets up tenant context for testing."""
-    token = CURRENT_TENANT_ID_CONTEXTVAR.set(TEST_TENANT_ID)
-    try:
-        yield  # Test runs here.
-    finally:
-        # Reset the tenant context after the test
-        CURRENT_TENANT_ID_CONTEXTVAR.reset(token)
-
-
-@pytest.fixture(scope="module")
-def httpx_client() -> Generator[httpx.Client, None, None]:
-    client = get_vespa_http_client()
-    try:
-        yield client
-    finally:
-        client.close()
-
-
-@pytest.fixture(scope="module")
-def vespa_document_index(
-    httpx_client: httpx.Client,
-    tenant_context: None,  # noqa: ARG001
-    test_index_name: str,
-) -> Generator[VespaIndex, None, None]:
-    vespa_index = VespaIndex(
-        index_name=test_index_name,
-        secondary_index_name=None,
-        large_chunks_enabled=False,
-        secondary_large_chunks_enabled=None,
-        multitenant=MULTI_TENANT,
-        httpx_client=httpx_client,
-    )
-    backend_dir = os.path.abspath(
-        os.path.join(os.path.dirname(__file__), "..", "..", "..")
-    )
-    with patch("os.getcwd", return_value=backend_dir):
-        vespa_index.ensure_indices_exist(
-            primary_embedding_dim=128,
-            primary_embedding_precision=EmbeddingPrecision.FLOAT,
-            secondary_index_embedding_dim=None,
-            secondary_index_embedding_precision=None,
-        )
-    # Verify Vespa is running, fails the test if not. Try 90 seconds for testing
-    # in CI. We have to do this here because this endpoint only becomes live
-    # once we create an index.
-    if not wait_for_vespa_with_timeout(wait_limit=90):
-        pytest.fail("Vespa is not available.")
-
-    # Wait until the schema is actually ready for writes on content nodes. We
-    # probe by attempting a PUT; 200 means the schema is live, 400 means not
-    # yet. This is so scuffed but running the test is really flakey otherwise;
-    # this is only temporary until we entirely move off of Vespa.
-    probe_doc = {
-        "fields": {
-            "document_id": "__probe__",
-            "chunk_id": 0,
-            "blurb": "",
-            "title": "",
-            "skip_title": True,
-            "content": "",
-            "content_summary": "",
-            "source_type": "file",
-            "source_links": "null",
-            "semantic_identifier": "",
-            "section_continuation": False,
-            "large_chunk_reference_ids": [],
-            "metadata": "{}",
-            "metadata_list": [],
-            "metadata_suffix": "",
-            "chunk_context": "",
-            "doc_summary": "",
-            "embeddings": {"full_chunk": [1.0] + [0.0] * 127},
-            "access_control_list": {},
-            "document_sets": {},
-            "image_file_name": None,
-            "user_project": [],
-            "personas": [],
-            "boost": 0.0,
-            "aggregated_chunk_boost_factor": 0.0,
-            "primary_owners": [],
-            "secondary_owners": [],
-        }
-    }
-    schema_ready = False
-    probe_url = (
-        f"http://localhost:8081/document/v1/default/{test_index_name}/docid/__probe__"
-    )
-    for _ in range(60):
-        resp = httpx_client.post(probe_url, json=probe_doc)
-        if resp.status_code == 200:
-            schema_ready = True
-            # Clean up the probe document.
-            httpx_client.delete(probe_url)
-            break
-        time.sleep(1)
-    if not schema_ready:
-        pytest.fail(f"Vespa schema '{test_index_name}' did not become ready in time.")
-
-    yield vespa_index  # Test runs here.
-
-    # TODO(ENG-3765)(andrei): Explicitly cleanup index. Not immediately
-    # pressing; in CI we should be using fresh instances of dependencies each
-    # time anyway.
-
-
-@pytest.fixture(scope="module")
-def opensearch_document_index(
-    opensearch_available: None,  # noqa: ARG001
-    tenant_context: None,  # noqa: ARG001
-    test_index_name: str,
-) -> Generator[OpenSearchOldDocumentIndex, None, None]:
-    opensearch_index = OpenSearchOldDocumentIndex(
-        index_name=test_index_name,
-        embedding_dim=128,
-        embedding_precision=EmbeddingPrecision.FLOAT,
-        secondary_index_name=None,
-        secondary_embedding_dim=None,
-        secondary_embedding_precision=None,
-        large_chunks_enabled=False,
-        secondary_large_chunks_enabled=None,
-        multitenant=MULTI_TENANT,
-    )
-    opensearch_index.ensure_indices_exist(
-        primary_embedding_dim=128,
-        primary_embedding_precision=EmbeddingPrecision.FLOAT,
-        secondary_index_embedding_dim=None,
-        secondary_index_embedding_precision=None,
-    )
-
-    yield opensearch_index  # Test runs here.
-
-    # TODO(ENG-3765)(andrei): Explicitly cleanup index. Not immediately
-    # pressing; in CI we should be using fresh instances of dependencies each
-    # time anyway.
-
-
-@pytest.fixture(scope="module")
-def document_indices(
-    vespa_document_index: VespaIndex,
-    opensearch_document_index: OpenSearchOldDocumentIndex,
-) -> Generator[list[DocumentIndex], None, None]:
-    # Ideally these are parametrized; doing so with pytest fixtures is tricky.
-    yield [opensearch_document_index, vespa_document_index]  # Test runs here.
-
-
-@pytest.fixture(scope="function")
-def chunks(
-    tenant_context: None,  # noqa: ARG001
-) -> Generator[list[DocMetadataAwareIndexChunk], None, None]:
-    result = []
-    chunk_count = 5
-    doc_id = "test_doc"
-    tenant_id = get_current_tenant_id()
-    access = DocumentAccess.build(
-        user_emails=[],
-        user_groups=[],
-        external_user_emails=[],
-        external_user_group_ids=[],
-        is_public=True,
-    )
-    document_sets: set[str] = set()
-    user_project: list[int] = list()
-    personas: list[int] = list()
-    boost = 0
-    blurb = "blurb"
-    content = "content"
-    title_prefix = ""
-    doc_summary = ""
-    chunk_context = ""
-    title_embedding = [1.0] + [0] * 127
-    # Full 0 vectors are not supported for cos similarity.
-    embeddings = ChunkEmbedding(
-        full_embedding=[1.0] + [0] * 127, mini_chunk_embeddings=[]
-    )
-    source_document = Document(
-        id=doc_id,
-        semantic_identifier="semantic identifier",
-        source=DocumentSource.FILE,
-        sections=[],
-        metadata={},
-        title="title",
-    )
-    metadata_suffix_keyword = ""
-    image_file_id = None
-    source_links: dict[int, str] = {0: ""}
-    ancestor_hierarchy_node_ids: list[int] = []
-    for i in range(chunk_count):
-        result.append(
-            DocMetadataAwareIndexChunk(
-                tenant_id=tenant_id,
-                access=access,
-                document_sets=document_sets,
-                user_project=user_project,
-                personas=personas,
-                boost=boost,
-                aggregated_chunk_boost_factor=0,
-                ancestor_hierarchy_node_ids=ancestor_hierarchy_node_ids,
-                embeddings=embeddings,
-                title_embedding=title_embedding,
-                source_document=source_document,
-                title_prefix=title_prefix,
-                metadata_suffix_keyword=metadata_suffix_keyword,
-                metadata_suffix_semantic="",
-                contextual_rag_reserved_tokens=0,
-                doc_summary=doc_summary,
-                chunk_context=chunk_context,
-                mini_chunk_texts=None,
-                large_chunk_id=None,
-                chunk_id=i,
-                blurb=blurb,
-                content=content,
-                source_links=source_links,
-                image_file_id=image_file_id,
-                section_continuation=False,
-            )
-        )
-    yield result  # Test runs here.
-
-
-@pytest.fixture(scope="function")
-def index_batch_params(
-    tenant_context: None,  # noqa: ARG001
-) -> Generator[IndexBatchParams, None, None]:
-    # WARNING: doc_id_to_previous_chunk_cnt={"test_doc": 0} is hardcoded to 0,
-    # which is only correct on the very first index call. The document_indices
-    # fixture is scope="module", meaning the same OpenSearch and Vespa backends
-    # persist across all test functions in this module. When a second test
-    # function uses this fixture and calls document_index.index(...), the
-    # backend already has 5 chunks for "test_doc" from the previous test run,
-    # but the batch params still claim 0 prior chunks exist. This can lead to
-    # orphaned/duplicate chunks that make subsequent assertions incorrect.
-    # TODO: Whenever adding a second test, either change this or cleanup the
-    # index between test cases.
-    yield IndexBatchParams(
-        doc_id_to_previous_chunk_cnt={"test_doc": 0},
-        doc_id_to_new_chunk_cnt={"test_doc": 5},
-        tenant_id=get_current_tenant_id(),
-        large_chunks_enabled=False,
-    )
-
-
-class TestDocumentIndexOld:
-    """Tests the old DocumentIndex interface."""
-
-    def test_update_single_can_clear_user_projects_and_personas(
-        self,
-        document_indices: list[DocumentIndex],
-        # This test case assumes all these chunks correspond to one document.
-        chunks: list[DocMetadataAwareIndexChunk],
-        index_batch_params: IndexBatchParams,
-    ) -> None:
-        """
-        Tests that update_single can clear user_projects and personas.
-        """
-        for document_index in document_indices:
-            # Precondition.
-            # Ensure there is some non-empty value for user project and
-            # personas.
-            for chunk in chunks:
-                chunk.user_project = [1]
-                chunk.personas = [2]
-            document_index.index(chunks, index_batch_params)
-
-            # Ensure that we can get chunks as expected with filters.
-            doc_id = chunks[0].source_document.id
-            chunk_count = len(chunks)
-            tenant_id = get_current_tenant_id()
-            # We need to specify the chunk index range and specify
-            # batch_retrieval=True below to trigger the codepath for Vespa's
-            # search API, which uses the expected additive filtering for
-            # project_id and persona_id. Otherwise we would use the codepath for
-            # the visit API, which does not have this kind of filtering
-            # implemented.
-            chunk_request = VespaChunkRequest(
-                document_id=doc_id, min_chunk_ind=0, max_chunk_ind=chunk_count - 1
-            )
-            project_persona_filters = IndexFilters(
-                access_control_list=None,
-                tenant_id=tenant_id,
-                project_id=1,
-                persona_id=2,
-                # We need this even though none of the chunks belong to a
-                # document set because project_id and persona_id are only
-                # additive filters in the event the agent has knowledge scope;
-                # if the agent does not, it is implied that it can see
-                # everything it is allowed to.
-                document_set=["1"],
-            )
-            # Not best practice here but the API for refreshing the index to
-            # ensure that the latest data is present is not exposed in this
-            # class and is not the same for Vespa and OpenSearch, so we just
-            # tolerate a sleep for now. As a consequence the number of tests in
-            # this suite should be small. We only need to tolerate this for as
-            # long as we continue to use Vespa, we can consider exposing
-            # something for OpenSearch later.
-            time.sleep(1)
-            inference_chunks = document_index.id_based_retrieval(
-                chunk_requests=[chunk_request],
-                filters=project_persona_filters,
-                batch_retrieval=True,
-            )
-            assert len(inference_chunks) == chunk_count
-            # Sort by chunk id to easily test if we have all chunks.
-            for i, inference_chunk in enumerate(
-                sorted(inference_chunks, key=lambda x: x.chunk_id)
-            ):
-                assert inference_chunk.chunk_id == i
-                assert inference_chunk.document_id == doc_id
-
-            # Under test.
-            # Explicitly set empty fields here.
-            user_fields = VespaDocumentUserFields(user_projects=[], personas=[])
-            document_index.update_single(
-                doc_id=doc_id,
-                chunk_count=chunk_count,
-                tenant_id=tenant_id,
-                fields=None,
-                user_fields=user_fields,
-            )
-
-            # Postcondition.
-            filters = IndexFilters(access_control_list=None, tenant_id=tenant_id)
-            # We should expect to get back all expected chunks with no filters.
-            # Again, not best practice here.
-            time.sleep(1)
-            inference_chunks = document_index.id_based_retrieval(
-                chunk_requests=[chunk_request], filters=filters, batch_retrieval=True
-            )
-            assert len(inference_chunks) == chunk_count
-            # Sort by chunk id to easily test if we have all chunks.
-            for i, inference_chunk in enumerate(
-                sorted(inference_chunks, key=lambda x: x.chunk_id)
-            ):
-                assert inference_chunk.chunk_id == i
-                assert inference_chunk.document_id == doc_id
-            # Now, we should expect to not get any chunks if we specify the user
-            # project and personas filters.
-            inference_chunks = document_index.id_based_retrieval(
-                chunk_requests=[chunk_request],
-                filters=project_persona_filters,
-                batch_retrieval=True,
-            )
-            assert len(inference_chunks) == 0

backend/tests/external_dependency_unit/opensearch_migration/test_opensearch_migration_tasks.py

@@ -239,8 +239,6 @@ def full_deployment_setup() -> Generator[None, None, None]:
     NOTE: We deliberately duplicate this logic from
     backend/tests/external_dependency_unit/conftest.py because we need to set
     opensearch_available just for this module, not the entire test session.
-
-    TODO(ENG-3764)(andrei): Consolidate some of these test fixtures.
     """
     # Patch ENABLE_OPENSEARCH_INDEXING_FOR_ONYX just for this test because we
     # don't yet want that enabled for all tests.

backend/tests/unit/onyx/chat/test_multi_model_types.py

@@ -1,112 +0,0 @@
-"""Unit tests for multi-model schema and Pydantic model additions."""
-
-from datetime import datetime
-
-from onyx.configs.constants import MessageType
-from onyx.llm.override_models import LLMOverride
-from onyx.server.query_and_chat.models import ChatMessageDetail
-from onyx.server.query_and_chat.models import MultiModelMessageResponseIDInfo
-from onyx.server.query_and_chat.models import SendMessageRequest
-from onyx.server.query_and_chat.placement import Placement
-
-
-def test_placement_model_index_default_none() -> None:
-    p = Placement(turn_index=0)
-    assert p.model_index is None
-
-
-def test_placement_model_index_set() -> None:
-    p = Placement(turn_index=0, model_index=2)
-    assert p.model_index == 2
-
-
-def test_placement_serialization_with_model_index() -> None:
-    p = Placement(turn_index=1, tab_index=0, model_index=1)
-    data = p.model_dump()
-    assert data["model_index"] == 1
-    restored = Placement(**data)
-    assert restored.model_index == 1
-
-
-def test_multi_model_message_response_id_info() -> None:
-    info = MultiModelMessageResponseIDInfo(
-        user_message_id=42,
-        reserved_assistant_message_ids=[100, 101, 102],
-        model_names=["gpt-4", "claude-3-opus", "gemini-pro"],
-    )
-    data = info.model_dump()
-    assert data["user_message_id"] == 42
-    assert len(data["reserved_assistant_message_ids"]) == 3
-    assert len(data["model_names"]) == 3
-
-
-def test_multi_model_message_response_id_info_null_user() -> None:
-    info = MultiModelMessageResponseIDInfo(
-        user_message_id=None,
-        reserved_assistant_message_ids=[10],
-        model_names=["gpt-4"],
-    )
-    assert info.user_message_id is None
-
-
-def test_send_message_request_llm_overrides_none_by_default() -> None:
-    req = SendMessageRequest(
-        message="hello",
-        chat_session_id="00000000-0000-0000-0000-000000000001",
-    )
-    assert req.llm_overrides is None
-    assert req.llm_override is None
-
-
-def test_send_message_request_with_llm_overrides() -> None:
-    overrides = [
-        LLMOverride(model_provider="openai", model_version="gpt-4"),
-        LLMOverride(model_provider="anthropic", model_version="claude-3-opus"),
-    ]
-    req = SendMessageRequest(
-        message="compare these",
-        chat_session_id="00000000-0000-0000-0000-000000000001",
-        llm_overrides=overrides,
-    )
-    assert req.llm_overrides is not None
-    assert len(req.llm_overrides) == 2
-
-
-def test_send_message_request_backward_compat_single_override() -> None:
-    """Existing single llm_override still works alongside new llm_overrides field."""
-    req = SendMessageRequest(
-        message="single model",
-        chat_session_id="00000000-0000-0000-0000-000000000001",
-        llm_override=LLMOverride(model_provider="openai", model_version="gpt-4"),
-    )
-    assert req.llm_override is not None
-    assert req.llm_overrides is None
-
-
-def test_chat_message_detail_multi_model_fields_default_none() -> None:
-    detail = ChatMessageDetail(
-        message_id=1,
-        message="hello",
-        message_type=MessageType.USER,
-        time_sent=datetime.now(),
-        files=[],
-    )
-    assert detail.preferred_response_id is None
-    assert detail.model_display_name is None
-
-
-def test_chat_message_detail_multi_model_fields_set() -> None:
-    detail = ChatMessageDetail(
-        message_id=1,
-        message="response from gpt-4",
-        message_type=MessageType.ASSISTANT,
-        time_sent=datetime.now(),
-        files=[],
-        preferred_response_id=42,
-        model_display_name="GPT-4",
-    )
-    assert detail.preferred_response_id == 42
-    assert detail.model_display_name == "GPT-4"
-    data = detail.model_dump()
-    assert data["preferred_response_id"] == 42
-    assert data["model_display_name"] == "GPT-4"

backend/tests/unit/onyx/connectors/sharepoint/test_delta_checkpointing.py

@@ -6,7 +6,6 @@ Validates that:
 - Crash + resume skips already-processed pages
 - BFS (folder-scoped) drives process all items in one call
 - 410 Gone triggers a full-resync URL in the checkpoint
-- Duplicate document IDs across delta pages are deduplicated
 """
 
 from __future__ import annotations
@@ -458,228 +457,3 @@ class TestDeltaPageFetchFailure:
         assert final_cp.current_drive_name is None
         assert final_cp.current_drive_id is None
         assert final_cp.current_drive_delta_next_link is None
-
-
-class TestDeltaDuplicateDocumentDedup:
-    """The Microsoft Graph delta API can return the same item on multiple
-    pages.  Documents already yielded should be skipped via
-    checkpoint.seen_document_ids."""
-
-    def test_duplicate_across_pages_is_skipped(
-        self, monkeypatch: pytest.MonkeyPatch
-    ) -> None:
-        """Item 'dup' appears on both page 1 and page 2.  It should only be
-        yielded once."""
-        connector = _setup_connector(monkeypatch)
-        _mock_convert(monkeypatch)
-
-        call_count = 0
-
-        def fake_fetch_page(
-            self: SharepointConnector,  # noqa: ARG001
-            page_url: str,  # noqa: ARG001
-            drive_id: str,  # noqa: ARG001
-            start: datetime | None = None,  # noqa: ARG001
-            end: datetime | None = None,  # noqa: ARG001
-            page_size: int = 200,  # noqa: ARG001
-        ) -> tuple[list[DriveItemData], str | None]:
-            nonlocal call_count
-            call_count += 1
-            if call_count == 1:
-                return [_make_item("a"), _make_item("dup")], "https://next2"
-            return [_make_item("dup"), _make_item("b")], None
-
-        monkeypatch.setattr(
-            SharepointConnector, "_fetch_one_delta_page", fake_fetch_page
-        )
-
-        checkpoint = _build_ready_checkpoint()
-
-        # Page 1: yields a, dup
-        gen = connector._load_from_checkpoint(
-            _START_TS, _END_TS, checkpoint, include_permissions=False
-        )
-        yielded, checkpoint = _consume_generator(gen)
-        docs = _docs_from(yielded)
-        assert [d.id for d in docs] == ["a", "dup"]
-        assert "dup" in checkpoint.seen_document_ids
-
-        # Page 2: dup should be skipped, only b yielded
-        gen = connector._load_from_checkpoint(
-            _START_TS, _END_TS, checkpoint, include_permissions=False
-        )
-        yielded, checkpoint = _consume_generator(gen)
-        docs = _docs_from(yielded)
-        assert [d.id for d in docs] == ["b"]
-
-    def test_duplicate_within_same_page_is_skipped(
-        self, monkeypatch: pytest.MonkeyPatch
-    ) -> None:
-        """If the same item appears twice on a single delta page, only the
-        first occurrence should be yielded."""
-        connector = _setup_connector(monkeypatch)
-        _mock_convert(monkeypatch)
-
-        def fake_fetch_page(
-            self: SharepointConnector,  # noqa: ARG001
-            page_url: str,  # noqa: ARG001
-            drive_id: str,  # noqa: ARG001
-            start: datetime | None = None,  # noqa: ARG001
-            end: datetime | None = None,  # noqa: ARG001
-            page_size: int = 200,  # noqa: ARG001
-        ) -> tuple[list[DriveItemData], str | None]:
-            return [_make_item("x"), _make_item("x"), _make_item("y")], None
-
-        monkeypatch.setattr(
-            SharepointConnector, "_fetch_one_delta_page", fake_fetch_page
-        )
-
-        checkpoint = _build_ready_checkpoint()
-        gen = connector._load_from_checkpoint(
-            _START_TS, _END_TS, checkpoint, include_permissions=False
-        )
-        yielded, checkpoint = _consume_generator(gen)
-        docs = _docs_from(yielded)
-        assert [d.id for d in docs] == ["x", "y"]
-
-    def test_seen_ids_survive_checkpoint_serialization(
-        self, monkeypatch: pytest.MonkeyPatch
-    ) -> None:
-        """seen_document_ids must survive JSON serialization so that
-        dedup works across crash + resume."""
-        connector = _setup_connector(monkeypatch)
-        _mock_convert(monkeypatch)
-
-        call_count = 0
-
-        def fake_fetch_page(
-            self: SharepointConnector,  # noqa: ARG001
-            page_url: str,  # noqa: ARG001
-            drive_id: str,  # noqa: ARG001
-            start: datetime | None = None,  # noqa: ARG001
-            end: datetime | None = None,  # noqa: ARG001
-            page_size: int = 200,  # noqa: ARG001
-        ) -> tuple[list[DriveItemData], str | None]:
-            nonlocal call_count
-            call_count += 1
-            if call_count == 1:
-                return [_make_item("a")], "https://next2"
-            return [_make_item("a"), _make_item("b")], None
-
-        monkeypatch.setattr(
-            SharepointConnector, "_fetch_one_delta_page", fake_fetch_page
-        )
-
-        checkpoint = _build_ready_checkpoint()
-
-        # Page 1
-        gen = connector._load_from_checkpoint(
-            _START_TS, _END_TS, checkpoint, include_permissions=False
-        )
-        _, checkpoint = _consume_generator(gen)
-        assert "a" in checkpoint.seen_document_ids
-
-        # Simulate crash: round-trip through JSON
-        restored = SharepointConnectorCheckpoint.model_validate_json(
-            checkpoint.model_dump_json()
-        )
-        assert "a" in restored.seen_document_ids
-
-        # Page 2 with restored checkpoint: 'a' should be skipped
-        connector2 = _setup_connector(monkeypatch)
-        _mock_convert(monkeypatch)
-        monkeypatch.setattr(
-            SharepointConnector, "_fetch_one_delta_page", fake_fetch_page
-        )
-
-        gen = connector2._load_from_checkpoint(
-            _START_TS, _END_TS, restored, include_permissions=False
-        )
-        yielded, final_cp = _consume_generator(gen)
-        docs = _docs_from(yielded)
-        assert [d.id for d in docs] == ["b"]
-
-    def test_no_dedup_across_separate_indexing_runs(
-        self, monkeypatch: pytest.MonkeyPatch
-    ) -> None:
-        """A fresh checkpoint (new indexing run) should have an empty
-        seen_document_ids, so previously-indexed docs are re-processed."""
-        connector = _setup_connector(monkeypatch)
-        _mock_convert(monkeypatch)
-
-        def fake_fetch_page(
-            self: SharepointConnector,  # noqa: ARG001
-            page_url: str,  # noqa: ARG001
-            drive_id: str,  # noqa: ARG001
-            start: datetime | None = None,  # noqa: ARG001
-            end: datetime | None = None,  # noqa: ARG001
-            page_size: int = 200,  # noqa: ARG001
-        ) -> tuple[list[DriveItemData], str | None]:
-            return [_make_item("a")], None
-
-        monkeypatch.setattr(
-            SharepointConnector, "_fetch_one_delta_page", fake_fetch_page
-        )
-
-        # First run
-        cp1 = _build_ready_checkpoint()
-        gen = connector._load_from_checkpoint(
-            _START_TS, _END_TS, cp1, include_permissions=False
-        )
-        yielded, _ = _consume_generator(gen)
-        assert len(_docs_from(yielded)) == 1
-
-        # Second run with a fresh checkpoint — same doc should appear again
-        cp2 = _build_ready_checkpoint()
-        assert len(cp2.seen_document_ids) == 0
-        gen = connector._load_from_checkpoint(
-            _START_TS, _END_TS, cp2, include_permissions=False
-        )
-        yielded, _ = _consume_generator(gen)
-        assert len(_docs_from(yielded)) == 1
-
-    def test_same_id_across_drives_not_skipped(
-        self, monkeypatch: pytest.MonkeyPatch
-    ) -> None:
-        """Graph item IDs are only unique within a drive.  An item in drive B
-        that happens to share an ID with an item already seen in drive A must
-        NOT be skipped."""
-        connector = _setup_connector(monkeypatch)
-        _mock_convert(monkeypatch)
-
-        def fake_fetch_page(
-            self: SharepointConnector,  # noqa: ARG001
-            page_url: str,  # noqa: ARG001
-            drive_id: str,  # noqa: ARG001
-            start: datetime | None = None,  # noqa: ARG001
-            end: datetime | None = None,  # noqa: ARG001
-            page_size: int = 200,  # noqa: ARG001
-        ) -> tuple[list[DriveItemData], str | None]:
-            return [_make_item("shared-id")], None
-
-        monkeypatch.setattr(
-            SharepointConnector, "_fetch_one_delta_page", fake_fetch_page
-        )
-
-        checkpoint = _build_ready_checkpoint(drive_names=["DriveA", "DriveB"])
-
-        # Drive A: yields the item
-        gen = connector._load_from_checkpoint(
-            _START_TS, _END_TS, checkpoint, include_permissions=False
-        )
-        yielded, checkpoint = _consume_generator(gen)
-        docs = _docs_from(yielded)
-        assert len(docs) == 1
-        assert docs[0].id == "shared-id"
-
-        # seen_document_ids should have been cleared when drive A finished
-        assert len(checkpoint.seen_document_ids) == 0
-
-        # Drive B: same ID must be yielded again (different drive)
-        gen = connector._load_from_checkpoint(
-            _START_TS, _END_TS, checkpoint, include_permissions=False
-        )
-        yielded, checkpoint = _consume_generator(gen)
-        docs = _docs_from(yielded)
-        assert len(docs) == 1
-        assert docs[0].id == "shared-id"

backend/tests/unit/onyx/db/test_llm_sync.py

@@ -7,7 +7,6 @@ import pytest
 
 from onyx.db.llm import sync_model_configurations
 from onyx.llm.constants import LlmProviderNames
-from onyx.server.manage.llm.models import SyncModelEntry
 
 
 class TestSyncModelConfigurations:
@@ -26,18 +25,18 @@ class TestSyncModelConfigurations:
             "onyx.db.llm.fetch_existing_llm_provider", return_value=mock_provider
         ):
             models = [
-                SyncModelEntry(
-                    name="gpt-4",
-                    display_name="GPT-4",
-                    max_input_tokens=128000,
-                    supports_image_input=True,
-                ),
-                SyncModelEntry(
-                    name="gpt-4o",
-                    display_name="GPT-4o",
-                    max_input_tokens=128000,
-                    supports_image_input=True,
-                ),
+                {
+                    "name": "gpt-4",
+                    "display_name": "GPT-4",
+                    "max_input_tokens": 128000,
+                    "supports_image_input": True,
+                },
+                {
+                    "name": "gpt-4o",
+                    "display_name": "GPT-4o",
+                    "max_input_tokens": 128000,
+                    "supports_image_input": True,
+                },
             ]
 
             result = sync_model_configurations(
@@ -68,18 +67,18 @@ class TestSyncModelConfigurations:
             "onyx.db.llm.fetch_existing_llm_provider", return_value=mock_provider
         ):
             models = [
-                SyncModelEntry(
-                    name="gpt-4",  # Existing - should be skipped
-                    display_name="GPT-4",
-                    max_input_tokens=128000,
-                    supports_image_input=True,
-                ),
-                SyncModelEntry(
-                    name="gpt-4o",  # New - should be inserted
-                    display_name="GPT-4o",
-                    max_input_tokens=128000,
-                    supports_image_input=True,
-                ),
+                {
+                    "name": "gpt-4",  # Existing - should be skipped
+                    "display_name": "GPT-4",
+                    "max_input_tokens": 128000,
+                    "supports_image_input": True,
+                },
+                {
+                    "name": "gpt-4o",  # New - should be inserted
+                    "display_name": "GPT-4o",
+                    "max_input_tokens": 128000,
+                    "supports_image_input": True,
+                },
             ]
 
             result = sync_model_configurations(
@@ -106,12 +105,12 @@ class TestSyncModelConfigurations:
             "onyx.db.llm.fetch_existing_llm_provider", return_value=mock_provider
         ):
             models = [
-                SyncModelEntry(
-                    name="gpt-4",  # Already exists
-                    display_name="GPT-4",
-                    max_input_tokens=128000,
-                    supports_image_input=True,
-                ),
+                {
+                    "name": "gpt-4",  # Already exists
+                    "display_name": "GPT-4",
+                    "max_input_tokens": 128000,
+                    "supports_image_input": True,
+                },
             ]
 
             result = sync_model_configurations(
@@ -132,7 +131,7 @@ class TestSyncModelConfigurations:
                 sync_model_configurations(
                     db_session=mock_session,
                     provider_name="nonexistent",
-                    models=[SyncModelEntry(name="model", display_name="Model")],
+                    models=[{"name": "model", "display_name": "Model"}],
                 )
 
     def test_handles_missing_optional_fields(self) -> None:
@@ -146,12 +145,12 @@ class TestSyncModelConfigurations:
         with patch(
             "onyx.db.llm.fetch_existing_llm_provider", return_value=mock_provider
         ):
-            # Model with only required fields (max_input_tokens and supports_image_input default)
+            # Model with only required fields
             models = [
-                SyncModelEntry(
-                    name="model-1",
-                    display_name="Model 1",
-                ),
+                {
+                    "name": "model-1",
+                    # No display_name, max_input_tokens, or supports_image_input
+                },
             ]
 
             result = sync_model_configurations(

backend/tests/unit/onyx/llm/test_true_openai_model.py

@@ -26,6 +26,14 @@ class TestIsTrueOpenAIModel:
         """Test that real OpenAI GPT-4o-mini model is correctly identified."""
         assert is_true_openai_model(LlmProviderNames.OPENAI, "gpt-4o-mini") is True
 
+    def test_real_openai_o1_preview(self) -> None:
+        """Test that real OpenAI o1-preview reasoning model is correctly identified."""
+        assert is_true_openai_model(LlmProviderNames.OPENAI, "o1-preview") is True
+
+    def test_real_openai_o1_mini(self) -> None:
+        """Test that real OpenAI o1-mini reasoning model is correctly identified."""
+        assert is_true_openai_model(LlmProviderNames.OPENAI, "o1-mini") is True
+
     def test_openai_with_provider_prefix(self) -> None:
         """Test that OpenAI model with provider prefix is correctly identified."""
         assert is_true_openai_model(LlmProviderNames.OPENAI, "openai/gpt-4") is False

backend/tests/unit/onyx/onyxbot/test_handle_regular_answer.py

@@ -1,204 +0,0 @@
-"""Tests for Slack channel reference resolution and tag filtering
-in handle_regular_answer.py."""
-
-from unittest.mock import MagicMock
-
-from slack_sdk.errors import SlackApiError
-
-from onyx.context.search.models import Tag
-from onyx.onyxbot.slack.constants import SLACK_CHANNEL_REF_PATTERN
-from onyx.onyxbot.slack.handlers.handle_regular_answer import resolve_channel_references
-
-
-# ---------------------------------------------------------------------------
-# Helpers
-# ---------------------------------------------------------------------------
-
-
-def _mock_client_with_channels(
-    channel_map: dict[str, str],
-) -> MagicMock:
-    """Return a mock WebClient where conversations_info resolves IDs to names."""
-    client = MagicMock()
-
-    def _conversations_info(channel: str) -> MagicMock:
-        if channel in channel_map:
-            resp = MagicMock()
-            resp.validate = MagicMock()
-            resp.__getitem__ = lambda _self, key: {
-                "channel": {
-                    "name": channel_map[channel],
-                    "is_im": False,
-                    "is_mpim": False,
-                }
-            }[key]
-            return resp
-        raise SlackApiError("channel_not_found", response=MagicMock())
-
-    client.conversations_info = _conversations_info
-    return client
-
-
-def _mock_logger() -> MagicMock:
-    return MagicMock()
-
-
-# ---------------------------------------------------------------------------
-# SLACK_CHANNEL_REF_PATTERN regex tests
-# ---------------------------------------------------------------------------
-
-
-class TestSlackChannelRefPattern:
-    def test_matches_bare_channel_id(self) -> None:
-        matches = SLACK_CHANNEL_REF_PATTERN.findall("<#C097NBWMY8Y>")
-        assert matches == [("C097NBWMY8Y", "")]
-
-    def test_matches_channel_id_with_name(self) -> None:
-        matches = SLACK_CHANNEL_REF_PATTERN.findall("<#C097NBWMY8Y|eng-infra>")
-        assert matches == [("C097NBWMY8Y", "eng-infra")]
-
-    def test_matches_multiple_channels(self) -> None:
-        msg = "compare <#C111AAA> and <#C222BBB|general>"
-        matches = SLACK_CHANNEL_REF_PATTERN.findall(msg)
-        assert len(matches) == 2
-        assert ("C111AAA", "") in matches
-        assert ("C222BBB", "general") in matches
-
-    def test_no_match_on_plain_text(self) -> None:
-        matches = SLACK_CHANNEL_REF_PATTERN.findall("no channels here")
-        assert matches == []
-
-    def test_no_match_on_user_mention(self) -> None:
-        matches = SLACK_CHANNEL_REF_PATTERN.findall("<@U12345>")
-        assert matches == []
-
-
-# ---------------------------------------------------------------------------
-# resolve_channel_references tests
-# ---------------------------------------------------------------------------
-
-
-class TestResolveChannelReferences:
-    def test_resolves_bare_channel_id_via_api(self) -> None:
-        client = _mock_client_with_channels({"C097NBWMY8Y": "eng-infra"})
-        logger = _mock_logger()
-
-        message, tags = resolve_channel_references(
-            message="summary of <#C097NBWMY8Y> this week",
-            client=client,
-            logger=logger,
-        )
-
-        assert message == "summary of #eng-infra this week"
-        assert len(tags) == 1
-        assert tags[0] == Tag(tag_key="Channel", tag_value="eng-infra")
-
-    def test_uses_name_from_pipe_format_without_api_call(self) -> None:
-        client = MagicMock()
-        logger = _mock_logger()
-
-        message, tags = resolve_channel_references(
-            message="check <#C097NBWMY8Y|eng-infra> for updates",
-            client=client,
-            logger=logger,
-        )
-
-        assert message == "check #eng-infra for updates"
-        assert tags == [Tag(tag_key="Channel", tag_value="eng-infra")]
-        # Should NOT have called the API since name was in the markup
-        client.conversations_info.assert_not_called()
-
-    def test_multiple_channels(self) -> None:
-        client = _mock_client_with_channels(
-            {
-                "C111AAA": "eng-infra",
-                "C222BBB": "eng-general",
-            }
-        )
-        logger = _mock_logger()
-
-        message, tags = resolve_channel_references(
-            message="compare <#C111AAA> and <#C222BBB>",
-            client=client,
-            logger=logger,
-        )
-
-        assert "#eng-infra" in message
-        assert "#eng-general" in message
-        assert "<#" not in message
-        assert len(tags) == 2
-        tag_values = {t.tag_value for t in tags}
-        assert tag_values == {"eng-infra", "eng-general"}
-
-    def test_no_channel_references_returns_unchanged(self) -> None:
-        client = MagicMock()
-        logger = _mock_logger()
-
-        message, tags = resolve_channel_references(
-            message="just a normal message with no channels",
-            client=client,
-            logger=logger,
-        )
-
-        assert message == "just a normal message with no channels"
-        assert tags == []
-
-    def test_api_failure_skips_channel_gracefully(self) -> None:
-        # Client that fails for all channel lookups
-        client = _mock_client_with_channels({})
-        logger = _mock_logger()
-
-        message, tags = resolve_channel_references(
-            message="check <#CBADID123>",
-            client=client,
-            logger=logger,
-        )
-
-        # Message should remain unchanged for the failed channel
-        assert "<#CBADID123>" in message
-        assert tags == []
-        logger.warning.assert_called_once()
-
-    def test_partial_failure_resolves_what_it_can(self) -> None:
-        # Only one of two channels resolves
-        client = _mock_client_with_channels({"C111AAA": "eng-infra"})
-        logger = _mock_logger()
-
-        message, tags = resolve_channel_references(
-            message="compare <#C111AAA> and <#CBADID123>",
-            client=client,
-            logger=logger,
-        )
-
-        assert "#eng-infra" in message
-        assert "<#CBADID123>" in message  # failed one stays raw
-        assert len(tags) == 1
-        assert tags[0].tag_value == "eng-infra"
-
-    def test_duplicate_channel_produces_single_tag(self) -> None:
-        client = _mock_client_with_channels({"C111AAA": "eng-infra"})
-        logger = _mock_logger()
-
-        message, tags = resolve_channel_references(
-            message="summarize <#C111AAA> and compare with <#C111AAA>",
-            client=client,
-            logger=logger,
-        )
-
-        assert message == "summarize #eng-infra and compare with #eng-infra"
-        assert len(tags) == 1
-        assert tags[0].tag_value == "eng-infra"
-
-    def test_mixed_pipe_and_bare_formats(self) -> None:
-        client = _mock_client_with_channels({"C222BBB": "random"})
-        logger = _mock_logger()
-
-        message, tags = resolve_channel_references(
-            message="see <#C111AAA|eng-infra> and <#C222BBB>",
-            client=client,
-            logger=logger,
-        )
-
-        assert "#eng-infra" in message
-        assert "#random" in message
-        assert len(tags) == 2

backend/tests/unit/onyx/server/manage/llm/test_fetch_models_api.py

@@ -1,19 +1,15 @@
 """Tests for LLM model fetch endpoints.
 
 These tests verify the full request/response flow for fetching models
-from dynamic providers (Ollama, OpenRouter, Litellm), including the
+from dynamic providers (Ollama, OpenRouter), including the
 sync-to-DB behavior when provider_name is specified.
 """
 
 from unittest.mock import MagicMock
 from unittest.mock import patch
 
-import httpx
 import pytest
 
-from onyx.error_handling.exceptions import OnyxError
-from onyx.server.manage.llm.models import LitellmFinalModelResponse
-from onyx.server.manage.llm.models import LitellmModelsRequest
 from onyx.server.manage.llm.models import LMStudioFinalModelResponse
 from onyx.server.manage.llm.models import LMStudioModelsRequest
 from onyx.server.manage.llm.models import OllamaFinalModelResponse
@@ -618,283 +614,3 @@ class TestGetLMStudioAvailableModels:
             request = LMStudioModelsRequest(api_base="http://localhost:1234")
             with pytest.raises(OnyxError):
                 get_lm_studio_available_models(request, MagicMock(), mock_session)
-
-
-class TestGetLitellmAvailableModels:
-    """Tests for the Litellm proxy model fetch endpoint."""
-
-    @pytest.fixture
-    def mock_litellm_response(self) -> dict:
-        """Mock response from Litellm /v1/models endpoint."""
-        return {
-            "data": [
-                {
-                    "id": "gpt-4o",
-                    "object": "model",
-                    "created": 1700000000,
-                    "owned_by": "openai",
-                },
-                {
-                    "id": "claude-3-5-sonnet",
-                    "object": "model",
-                    "created": 1700000001,
-                    "owned_by": "anthropic",
-                },
-                {
-                    "id": "gemini-pro",
-                    "object": "model",
-                    "created": 1700000002,
-                    "owned_by": "google",
-                },
-            ]
-        }
-
-    def test_returns_model_list(self, mock_litellm_response: dict) -> None:
-        """Test that endpoint returns properly formatted model list."""
-        from onyx.server.manage.llm.api import get_litellm_available_models
-
-        mock_session = MagicMock()
-
-        with patch("onyx.server.manage.llm.api.httpx.get") as mock_get:
-            mock_response = MagicMock()
-            mock_response.json.return_value = mock_litellm_response
-            mock_response.raise_for_status = MagicMock()
-            mock_get.return_value = mock_response
-
-            request = LitellmModelsRequest(
-                api_base="http://localhost:4000",
-                api_key="test-key",
-            )
-            results = get_litellm_available_models(request, MagicMock(), mock_session)
-
-            assert len(results) == 3
-            assert all(isinstance(r, LitellmFinalModelResponse) for r in results)
-
-    def test_model_fields_parsed_correctly(self, mock_litellm_response: dict) -> None:
-        """Test that provider_name and model_name are correctly extracted."""
-        from onyx.server.manage.llm.api import get_litellm_available_models
-
-        mock_session = MagicMock()
-
-        with patch("onyx.server.manage.llm.api.httpx.get") as mock_get:
-            mock_response = MagicMock()
-            mock_response.json.return_value = mock_litellm_response
-            mock_response.raise_for_status = MagicMock()
-            mock_get.return_value = mock_response
-
-            request = LitellmModelsRequest(
-                api_base="http://localhost:4000",
-                api_key="test-key",
-            )
-            results = get_litellm_available_models(request, MagicMock(), mock_session)
-
-            gpt = next(r for r in results if r.model_name == "gpt-4o")
-            assert gpt.provider_name == "openai"
-
-            claude = next(r for r in results if r.model_name == "claude-3-5-sonnet")
-            assert claude.provider_name == "anthropic"
-
-    def test_results_sorted_by_model_name(self, mock_litellm_response: dict) -> None:
-        """Test that results are alphabetically sorted by model_name."""
-        from onyx.server.manage.llm.api import get_litellm_available_models
-
-        mock_session = MagicMock()
-
-        with patch("onyx.server.manage.llm.api.httpx.get") as mock_get:
-            mock_response = MagicMock()
-            mock_response.json.return_value = mock_litellm_response
-            mock_response.raise_for_status = MagicMock()
-            mock_get.return_value = mock_response
-
-            request = LitellmModelsRequest(
-                api_base="http://localhost:4000",
-                api_key="test-key",
-            )
-            results = get_litellm_available_models(request, MagicMock(), mock_session)
-
-            model_names = [r.model_name for r in results]
-            assert model_names == sorted(model_names, key=str.lower)
-
-    def test_empty_data_raises_onyx_error(self) -> None:
-        """Test that empty model list raises OnyxError."""
-        from onyx.server.manage.llm.api import get_litellm_available_models
-
-        mock_session = MagicMock()
-
-        with patch("onyx.server.manage.llm.api.httpx.get") as mock_get:
-            mock_response = MagicMock()
-            mock_response.json.return_value = {"data": []}
-            mock_response.raise_for_status = MagicMock()
-            mock_get.return_value = mock_response
-
-            request = LitellmModelsRequest(
-                api_base="http://localhost:4000",
-                api_key="test-key",
-            )
-            with pytest.raises(OnyxError, match="No models found"):
-                get_litellm_available_models(request, MagicMock(), mock_session)
-
-    def test_missing_data_key_raises_onyx_error(self) -> None:
-        """Test that response without 'data' key raises OnyxError."""
-        from onyx.server.manage.llm.api import get_litellm_available_models
-
-        mock_session = MagicMock()
-
-        with patch("onyx.server.manage.llm.api.httpx.get") as mock_get:
-            mock_response = MagicMock()
-            mock_response.json.return_value = {}
-            mock_response.raise_for_status = MagicMock()
-            mock_get.return_value = mock_response
-
-            request = LitellmModelsRequest(
-                api_base="http://localhost:4000",
-                api_key="test-key",
-            )
-            with pytest.raises(OnyxError):
-                get_litellm_available_models(request, MagicMock(), mock_session)
-
-    def test_skips_unparseable_entries(self) -> None:
-        """Test that malformed model entries are skipped without failing."""
-        from onyx.server.manage.llm.api import get_litellm_available_models
-
-        mock_session = MagicMock()
-        response_with_bad_entry = {
-            "data": [
-                {
-                    "id": "gpt-4o",
-                    "object": "model",
-                    "created": 1700000000,
-                    "owned_by": "openai",
-                },
-                # Missing required fields
-                {"bad_field": "bad_value"},
-            ]
-        }
-
-        with patch("onyx.server.manage.llm.api.httpx.get") as mock_get:
-            mock_response = MagicMock()
-            mock_response.json.return_value = response_with_bad_entry
-            mock_response.raise_for_status = MagicMock()
-            mock_get.return_value = mock_response
-
-            request = LitellmModelsRequest(
-                api_base="http://localhost:4000",
-                api_key="test-key",
-            )
-            results = get_litellm_available_models(request, MagicMock(), mock_session)
-
-            assert len(results) == 1
-            assert results[0].model_name == "gpt-4o"
-
-    def test_all_entries_unparseable_raises_onyx_error(self) -> None:
-        """Test that OnyxError is raised when all entries fail to parse."""
-        from onyx.server.manage.llm.api import get_litellm_available_models
-
-        mock_session = MagicMock()
-        response_all_bad = {
-            "data": [
-                {"bad_field": "bad_value"},
-                {"another_bad": 123},
-            ]
-        }
-
-        with patch("onyx.server.manage.llm.api.httpx.get") as mock_get:
-            mock_response = MagicMock()
-            mock_response.json.return_value = response_all_bad
-            mock_response.raise_for_status = MagicMock()
-            mock_get.return_value = mock_response
-
-            request = LitellmModelsRequest(
-                api_base="http://localhost:4000",
-                api_key="test-key",
-            )
-            with pytest.raises(OnyxError, match="No compatible models"):
-                get_litellm_available_models(request, MagicMock(), mock_session)
-
-    def test_api_base_trailing_slash_handled(self) -> None:
-        """Test that trailing slashes in api_base are handled correctly."""
-        from onyx.server.manage.llm.api import get_litellm_available_models
-
-        mock_session = MagicMock()
-        mock_litellm_response = {
-            "data": [
-                {
-                    "id": "gpt-4o",
-                    "object": "model",
-                    "created": 1700000000,
-                    "owned_by": "openai",
-                },
-            ]
-        }
-
-        with patch("onyx.server.manage.llm.api.httpx.get") as mock_get:
-            mock_response = MagicMock()
-            mock_response.json.return_value = mock_litellm_response
-            mock_response.raise_for_status = MagicMock()
-            mock_get.return_value = mock_response
-
-            request = LitellmModelsRequest(
-                api_base="http://localhost:4000/",
-                api_key="test-key",
-            )
-            get_litellm_available_models(request, MagicMock(), mock_session)
-
-            # Should call /v1/models without double slashes
-            call_args = mock_get.call_args
-            assert call_args[0][0] == "http://localhost:4000/v1/models"
-
-    def test_connection_failure_raises_onyx_error(self) -> None:
-        """Test that connection failures are wrapped in OnyxError."""
-        from onyx.server.manage.llm.api import get_litellm_available_models
-
-        mock_session = MagicMock()
-
-        with patch("onyx.server.manage.llm.api.httpx.get") as mock_get:
-            mock_get.side_effect = Exception("Connection refused")
-
-            request = LitellmModelsRequest(
-                api_base="http://localhost:4000",
-                api_key="test-key",
-            )
-            with pytest.raises(OnyxError, match="Failed to fetch LiteLLM models"):
-                get_litellm_available_models(request, MagicMock(), mock_session)
-
-    def test_401_raises_authentication_error(self) -> None:
-        """Test that a 401 response raises OnyxError with authentication message."""
-        from onyx.server.manage.llm.api import get_litellm_available_models
-
-        mock_session = MagicMock()
-
-        with patch("onyx.server.manage.llm.api.httpx.get") as mock_get:
-            mock_response = MagicMock()
-            mock_response.status_code = 401
-            mock_get.side_effect = httpx.HTTPStatusError(
-                "Unauthorized", request=MagicMock(), response=mock_response
-            )
-
-            request = LitellmModelsRequest(
-                api_base="http://localhost:4000",
-                api_key="bad-key",
-            )
-            with pytest.raises(OnyxError, match="Authentication failed"):
-                get_litellm_available_models(request, MagicMock(), mock_session)
-
-    def test_404_raises_not_found_error(self) -> None:
-        """Test that a 404 response raises OnyxError with endpoint not found message."""
-        from onyx.server.manage.llm.api import get_litellm_available_models
-
-        mock_session = MagicMock()
-
-        with patch("onyx.server.manage.llm.api.httpx.get") as mock_get:
-            mock_response = MagicMock()
-            mock_response.status_code = 404
-            mock_get.side_effect = httpx.HTTPStatusError(
-                "Not Found", request=MagicMock(), response=mock_response
-            )
-
-            request = LitellmModelsRequest(
-                api_base="http://localhost:4000",
-                api_key="test-key",
-            )
-            with pytest.raises(OnyxError, match="endpoint not found"):
-                get_litellm_available_models(request, MagicMock(), mock_session)

backend/tests/unit/onyx/server/test_full_user_snapshot.py

@@ -1,54 +0,0 @@
-import datetime
-from unittest.mock import MagicMock
-from uuid import uuid4
-
-from onyx.auth.schemas import UserRole
-from onyx.server.models import FullUserSnapshot
-from onyx.server.models import UserGroupInfo
-
-
-def _mock_user(
-    personal_name: str | None = "Test User",
-    created_at: datetime.datetime | None = None,
-    updated_at: datetime.datetime | None = None,
-) -> MagicMock:
-    user = MagicMock()
-    user.id = uuid4()
-    user.email = "test@example.com"
-    user.role = UserRole.BASIC
-    user.is_active = True
-    user.password_configured = True
-    user.personal_name = personal_name
-    user.created_at = created_at or datetime.datetime(
-        2025, 1, 1, tzinfo=datetime.timezone.utc
-    )
-    user.updated_at = updated_at or datetime.datetime(
-        2025, 6, 15, tzinfo=datetime.timezone.utc
-    )
-    return user
-
-
-def test_from_user_model_includes_new_fields() -> None:
-    user = _mock_user(personal_name="Alice")
-    groups = [UserGroupInfo(id=1, name="Engineering")]
-
-    snapshot = FullUserSnapshot.from_user_model(user, groups=groups)
-
-    assert snapshot.personal_name == "Alice"
-    assert snapshot.created_at == user.created_at
-    assert snapshot.updated_at == user.updated_at
-    assert snapshot.groups == groups
-
-
-def test_from_user_model_defaults_groups_to_empty() -> None:
-    user = _mock_user()
-    snapshot = FullUserSnapshot.from_user_model(user)
-
-    assert snapshot.groups == []
-
-
-def test_from_user_model_personal_name_none() -> None:
-    user = _mock_user(personal_name=None)
-    snapshot = FullUserSnapshot.from_user_model(user)
-
-    assert snapshot.personal_name is None

deployment/docker_compose/docker-compose.dev.yml

@@ -38,11 +38,6 @@ services:
   opensearch:
     ports:
       - "9200:9200"
-    # Rootless Docker can reject the base OpenSearch ulimit settings, so clear
-    # the inherited block entirely in the dev override.
-    ulimits: !reset null
-    environment:
-      - bootstrap.memory_lock=false
 
   inference_model_server:
     ports:

deployment/docker_compose/docker-compose.onyx-lite.yml

@@ -15,8 +15,9 @@
 #                  -f docker-compose.dev.yml up -d --wait
 #
 # This overlay:
-#   - Moves Vespa (index), both model servers, code-interpreter, Redis (cache),
-#     and the background worker to profiles so they do not start by default
+#   - Moves Vespa (index), both model servers, OpenSearch, MinIO,
+#     Redis (cache), and the background worker to profiles so they do
+#     not start by default
 #   - Makes depends_on references to removed services optional
 #   - Sets DISABLE_VECTOR_DB=true on the api_server
 #   - Uses PostgreSQL for caching and auth instead of Redis
@@ -27,7 +28,8 @@
 #   --profile inference         Inference model server
 #   --profile background        Background worker (Celery) — also needs redis
 #   --profile redis             Redis cache
-#   --profile code-interpreter  Code interpreter
+#   --profile opensearch        OpenSearch
+#   --profile s3-filestore      MinIO (S3-compatible file store)
 # =============================================================================
 
 name: onyx
@@ -38,6 +40,9 @@ services:
       index:
         condition: service_started
         required: false
+      opensearch:
+        condition: service_started
+        required: false
       cache:
         condition: service_started
         required: false
@@ -84,4 +89,10 @@ services:
   inference_model_server:
     profiles: ["inference"]
 
-  code-interpreter: {}
+  # OpenSearch is not needed in lite mode (no indexing).
+  opensearch:
+    profiles: ["opensearch"]
+
+  # MinIO is not needed in lite mode (Postgres handles file storage).
+  minio:
+    profiles: ["s3-filestore"]

deployment/docker_compose/install.sh

@@ -1,8 +1,8 @@
 #!/bin/bash
 
-set -e
+set -euo pipefail
 
-# Expected resource requirements
+# Expected resource requirements (overridden below if --lite)
 EXPECTED_DOCKER_RAM_GB=10
 EXPECTED_DISK_GB=32
 
@@ -10,6 +10,11 @@ EXPECTED_DISK_GB=32
 SHUTDOWN_MODE=false
 DELETE_DATA_MODE=false
 INCLUDE_CRAFT=false  # Disabled by default, use --include-craft to enable
+LITE_MODE=false       # Disabled by default, use --lite to enable
+USE_LOCAL_FILES=false # Disabled by default, use --local to skip downloading config files
+NO_PROMPT=false
+DRY_RUN=false
+VERBOSE=false
 
 while [[ $# -gt 0 ]]; do
     case $1 in
@@ -25,6 +30,26 @@ while [[ $# -gt 0 ]]; do
             INCLUDE_CRAFT=true
             shift
             ;;
+        --lite)
+            LITE_MODE=true
+            shift
+            ;;
+        --local)
+            USE_LOCAL_FILES=true
+            shift
+            ;;
+        --no-prompt)
+            NO_PROMPT=true
+            shift
+            ;;
+        --dry-run)
+            DRY_RUN=true
+            shift
+            ;;
+        --verbose)
+            VERBOSE=true
+            shift
+            ;;
         --help|-h)
             echo "Onyx Installation Script"
             echo ""
@@ -32,15 +57,23 @@ while [[ $# -gt 0 ]]; do
             echo ""
             echo "Options:"
             echo "  --include-craft  Enable Onyx Craft (AI-powered web app building)"
+            echo "  --lite           Deploy Onyx Lite (no Vespa, Redis, or model servers)"
+            echo "  --local          Use existing config files instead of downloading from GitHub"
             echo "  --shutdown       Stop (pause) Onyx containers"
             echo "  --delete-data    Remove all Onyx data (containers, volumes, and files)"
+            echo "  --no-prompt      Run non-interactively with defaults (for CI/automation)"
+            echo "  --dry-run        Show what would be done without making changes"
+            echo "  --verbose        Show detailed output for debugging"
             echo "  --help, -h       Show this help message"
             echo ""
             echo "Examples:"
             echo "  $0                    # Install Onyx"
+            echo "  $0 --lite             # Install Onyx Lite (minimal deployment)"
             echo "  $0 --include-craft    # Install Onyx with Craft enabled"
             echo "  $0 --shutdown         # Pause Onyx services"
             echo "  $0 --delete-data      # Completely remove Onyx and all data"
+            echo "  $0 --local            # Re-run using existing config files on disk"
+            echo "  $0 --no-prompt        # Non-interactive install with defaults"
             exit 0
             ;;
         *)
@@ -51,8 +84,129 @@ while [[ $# -gt 0 ]]; do
     esac
 done
 
+if [[ "$VERBOSE" = true ]]; then
+    set -x
+fi
+
+if [[ "$LITE_MODE" = true ]] && [[ "$INCLUDE_CRAFT" = true ]]; then
+    echo "ERROR: --lite and --include-craft cannot be used together."
+    echo "Craft requires services (Vespa, Redis, background workers) that lite mode disables."
+    exit 1
+fi
+
+# When --lite is passed as a flag, lower resource thresholds early (before the
+# resource check). When lite is chosen interactively, the thresholds are adjusted
+# inside the new-deployment flow, after the resource check has already passed
+# with the standard thresholds — which is the safer direction.
+if [[ "$LITE_MODE" = true ]]; then
+    EXPECTED_DOCKER_RAM_GB=4
+    EXPECTED_DISK_GB=16
+fi
+
 INSTALL_ROOT="${INSTALL_PREFIX:-onyx_data}"
 
+LITE_COMPOSE_FILE="docker-compose.onyx-lite.yml"
+
+# Build the -f flags for docker compose.
+# Pass "true" as $1 to auto-detect a previously-downloaded lite overlay
+# (used by shutdown/delete-data so users don't need to remember --lite).
+# Without the argument, the lite overlay is only included when --lite was
+# explicitly passed — preventing install/start from silently staying in
+# lite mode just because the file exists on disk from a prior run.
+compose_file_args() {
+    local auto_detect="${1:-false}"
+    local args="-f docker-compose.yml"
+    if [[ "$LITE_MODE" = true ]] || { [[ "$auto_detect" = true ]] && [[ -f "${INSTALL_ROOT}/deployment/${LITE_COMPOSE_FILE}" ]]; }; then
+        args="$args -f ${LITE_COMPOSE_FILE}"
+    fi
+    echo "$args"
+}
+
+# --- Downloader detection (curl with wget fallback) ---
+DOWNLOADER=""
+detect_downloader() {
+    if command -v curl &> /dev/null; then
+        DOWNLOADER="curl"
+        return 0
+    fi
+    if command -v wget &> /dev/null; then
+        DOWNLOADER="wget"
+        return 0
+    fi
+    echo "ERROR: Neither curl nor wget found. Please install one and retry."
+    exit 1
+}
+detect_downloader
+
+download_file() {
+    local url="$1"
+    local output="$2"
+    if [[ "$DOWNLOADER" == "curl" ]]; then
+        curl -fsSL --retry 3 --retry-delay 2 --retry-connrefused -o "$output" "$url"
+    else
+        wget -q --tries=3 --timeout=20 -O "$output" "$url"
+    fi
+}
+
+# Ensures a required file is present. With --local, verifies the file exists on
+# disk. Otherwise, downloads it from the given URL. Returns 0 on success, 1 on
+# failure (caller should handle the exit).
+ensure_file() {
+    local path="$1"
+    local url="$2"
+    local desc="$3"
+
+    if [[ "$USE_LOCAL_FILES" = true ]]; then
+        if [[ -f "$path" ]]; then
+            print_success "Using existing ${desc}"
+            return 0
+        fi
+        print_error "Required file missing: ${desc} (${path})"
+        return 1
+    fi
+
+    print_info "Downloading ${desc}..."
+    if download_file "$url" "$path" 2>/dev/null; then
+        print_success "${desc} downloaded"
+        return 0
+    fi
+    print_error "Failed to download ${desc}"
+    print_info "Please ensure you have internet connection and try again"
+    return 1
+}
+
+# --- Interactive prompt helpers ---
+is_interactive() {
+    [[ "$NO_PROMPT" = false ]] && [[ -t 0 ]]
+}
+
+prompt_or_default() {
+    local prompt_text="$1"
+    local default_value="$2"
+    if is_interactive; then
+        read -p "$prompt_text" -r REPLY
+        if [[ -z "$REPLY" ]]; then
+            REPLY="$default_value"
+        fi
+    else
+        REPLY="$default_value"
+    fi
+}
+
+prompt_yn_or_default() {
+    local prompt_text="$1"
+    local default_value="$2"
+    if is_interactive; then
+        read -p "$prompt_text" -n 1 -r
+        echo ""
+        if [[ -z "$REPLY" ]]; then
+            REPLY="$default_value"
+        fi
+    else
+        REPLY="$default_value"
+    fi
+}
+
 # Colors for output
 RED='\033[0;31m'
 GREEN='\033[0;32m'
@@ -111,7 +265,7 @@ if [ "$SHUTDOWN_MODE" = true ]; then
             fi
 
             # Stop containers (without removing them)
-            (cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD -f docker-compose.yml stop)
+            (cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD $(compose_file_args true) stop)
             if [ $? -eq 0 ]; then
                 print_success "Onyx containers stopped (paused)"
             else
@@ -140,12 +294,17 @@ if [ "$DELETE_DATA_MODE" = true ]; then
     echo "  • All downloaded files and configurations"
     echo "  • All user data and documents"
     echo ""
-    read -p "Are you sure you want to continue? Type 'DELETE' to confirm: " -r
-    echo ""
-
-    if [ "$REPLY" != "DELETE" ]; then
-        print_info "Operation cancelled."
-        exit 0
+    if is_interactive; then
+        read -p "Are you sure you want to continue? Type 'DELETE' to confirm: " -r
+        echo ""
+        if [ "$REPLY" != "DELETE" ]; then
+            print_info "Operation cancelled."
+            exit 0
+        fi
+    else
+        print_error "Cannot confirm destructive operation in non-interactive mode."
+        print_info "Run interactively or remove the ${INSTALL_ROOT} directory manually."
+        exit 1
     fi
 
     print_info "Removing Onyx containers and volumes..."
@@ -164,7 +323,7 @@ if [ "$DELETE_DATA_MODE" = true ]; then
             fi
 
             # Stop and remove containers with volumes
-            (cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD -f docker-compose.yml down -v)
+            (cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD $(compose_file_args true) down -v)
             if [ $? -eq 0 ]; then
                 print_success "Onyx containers and volumes removed"
             else
@@ -186,6 +345,117 @@ if [ "$DELETE_DATA_MODE" = true ]; then
     exit 0
 fi
 
+# --- Auto-install Docker (Linux only) ---
+# Runs before the banner so a group-based re-exec doesn't repeat it.
+install_docker_linux() {
+    local distro_id=""
+    if [[ -f /etc/os-release ]]; then
+        distro_id="$(. /etc/os-release && echo "${ID:-}")"
+    fi
+
+    case "$distro_id" in
+        amzn)
+            print_info "Detected Amazon Linux — installing Docker via package manager..."
+            if command -v dnf &> /dev/null; then
+                sudo dnf install -y docker
+            else
+                sudo yum install -y docker
+            fi
+            ;;
+        *)
+            print_info "Installing Docker via get.docker.com..."
+            download_file "https://get.docker.com" /tmp/get-docker.sh
+            sudo sh /tmp/get-docker.sh
+            rm -f /tmp/get-docker.sh
+            ;;
+    esac
+
+    sudo systemctl start docker 2>/dev/null || sudo service docker start 2>/dev/null || true
+    sudo systemctl enable docker 2>/dev/null || true
+}
+
+# Detect OS (including WSL)
+IS_WSL=false
+if [[ -n "${WSL_DISTRO_NAME:-}" ]] || grep -qi microsoft /proc/version 2>/dev/null; then
+    IS_WSL=true
+fi
+
+# Dry-run: show plan and exit
+if [[ "$DRY_RUN" = true ]]; then
+    print_info "Dry run mode — showing what would happen:"
+    echo "  • Install root: ${INSTALL_ROOT}"
+    echo "  • Lite mode: ${LITE_MODE}"
+    echo "  • Include Craft: ${INCLUDE_CRAFT}"
+    echo "  • OS type: ${OSTYPE:-unknown} (WSL: ${IS_WSL})"
+    echo "  • Downloader: ${DOWNLOADER}"
+    echo ""
+    print_success "Dry run complete (no changes made)"
+    exit 0
+fi
+
+if ! command -v docker &> /dev/null; then
+    if [[ "$OSTYPE" == "linux-gnu"* ]] || [[ -n "${WSL_DISTRO_NAME:-}" ]]; then
+        install_docker_linux
+        if ! command -v docker &> /dev/null; then
+            print_error "Docker installation failed."
+            echo "  Visit: https://docs.docker.com/get-docker/"
+            exit 1
+        fi
+        print_success "Docker installed successfully"
+    fi
+fi
+
+# --- Auto-install Docker Compose plugin (Linux only) ---
+if command -v docker &> /dev/null \
+    && ! docker compose version &> /dev/null \
+    && ! command -v docker-compose &> /dev/null \
+    && { [[ "$OSTYPE" == "linux-gnu"* ]] || [[ -n "${WSL_DISTRO_NAME:-}" ]]; }; then
+
+    print_info "Docker Compose not found — installing plugin..."
+    COMPOSE_ARCH="$(uname -m)"
+    COMPOSE_URL="https://github.com/docker/compose/releases/latest/download/docker-compose-linux-${COMPOSE_ARCH}"
+    COMPOSE_DIR="/usr/local/lib/docker/cli-plugins"
+    COMPOSE_TMP="$(mktemp)"
+    sudo mkdir -p "$COMPOSE_DIR"
+    if download_file "$COMPOSE_URL" "$COMPOSE_TMP"; then
+        sudo mv "$COMPOSE_TMP" "$COMPOSE_DIR/docker-compose"
+        sudo chmod +x "$COMPOSE_DIR/docker-compose"
+        if docker compose version &> /dev/null; then
+            print_success "Docker Compose plugin installed"
+        else
+            print_error "Docker Compose plugin installed but not detected."
+            echo "  Visit: https://docs.docker.com/compose/install/"
+            exit 1
+        fi
+    else
+        rm -f "$COMPOSE_TMP"
+        print_error "Failed to download Docker Compose plugin."
+        echo "  Visit: https://docs.docker.com/compose/install/"
+        exit 1
+    fi
+fi
+
+# On Linux, ensure the current user can talk to the Docker daemon without
+# sudo.  If necessary, add them to the "docker" group and re-exec the
+# script under that group so the rest of the install proceeds normally.
+if command -v docker &> /dev/null \
+    && { [[ "$OSTYPE" == "linux-gnu"* ]] || [[ -n "${WSL_DISTRO_NAME:-}" ]]; } \
+    && [[ "$(id -u)" -ne 0 ]] \
+    && ! docker info &> /dev/null; then
+    if [[ "${_ONYX_REEXEC:-}" = "1" ]]; then
+        print_error "Cannot connect to Docker after group re-exec."
+        print_info "Log out and back in, then run the script again."
+        exit 1
+    fi
+    if ! getent group docker &> /dev/null; then
+        sudo groupadd docker
+    fi
+    print_info "Adding $USER to the docker group..."
+    sudo usermod -aG docker "$USER"
+    print_info "Re-launching with docker group active..."
+    exec sg docker -c "_ONYX_REEXEC=1 bash $(printf '%q ' "$0" "$@")"
+fi
+
 # ASCII Art Banner
 echo ""
 echo -e "${BLUE}${BOLD}"
@@ -209,8 +479,7 @@ echo "2. Check your system resources (Docker, memory, disk space)"
 echo "3. Guide you through deployment options (version, authentication)"
 echo ""
 
-# Only prompt for acknowledgment if running interactively
-if [ -t 0 ]; then
+if is_interactive; then
     echo -e "${YELLOW}${BOLD}Please acknowledge and press Enter to continue...${NC}"
     read -r
     echo ""
@@ -260,41 +529,35 @@ else
     exit 1
 fi
 
-# Function to compare version numbers
+# Returns 0 if $1 <= $2, 1 if $1 > $2
+# Handles missing or non-numeric parts gracefully (treats them as 0)
 version_compare() {
-    # Returns 0 if $1 <= $2, 1 if $1 > $2
-    local version1=$1
-    local version2=$2
+    local version1="${1:-0.0.0}"
+    local version2="${2:-0.0.0}"
 
-    # Split versions into components
-    local v1_major=$(echo $version1 | cut -d. -f1)
-    local v1_minor=$(echo $version1 | cut -d. -f2)
-    local v1_patch=$(echo $version1 | cut -d. -f3)
+    local v1_major v1_minor v1_patch v2_major v2_minor v2_patch
+    v1_major=$(echo "$version1" | cut -d. -f1)
+    v1_minor=$(echo "$version1" | cut -d. -f2)
+    v1_patch=$(echo "$version1" | cut -d. -f3)
+    v2_major=$(echo "$version2" | cut -d. -f1)
+    v2_minor=$(echo "$version2" | cut -d. -f2)
+    v2_patch=$(echo "$version2" | cut -d. -f3)
 
-    local v2_major=$(echo $version2 | cut -d. -f1)
-    local v2_minor=$(echo $version2 | cut -d. -f2)
-    local v2_patch=$(echo $version2 | cut -d. -f3)
+    # Default non-numeric or empty parts to 0
+    [[ "$v1_major" =~ ^[0-9]+$ ]] || v1_major=0
+    [[ "$v1_minor" =~ ^[0-9]+$ ]] || v1_minor=0
+    [[ "$v1_patch" =~ ^[0-9]+$ ]] || v1_patch=0
+    [[ "$v2_major" =~ ^[0-9]+$ ]] || v2_major=0
+    [[ "$v2_minor" =~ ^[0-9]+$ ]] || v2_minor=0
+    [[ "$v2_patch" =~ ^[0-9]+$ ]] || v2_patch=0
 
-    # Compare major version
-    if [ "$v1_major" -lt "$v2_major" ]; then
-        return 0
-    elif [ "$v1_major" -gt "$v2_major" ]; then
-        return 1
-    fi
+    if [ "$v1_major" -lt "$v2_major" ]; then return 0
+    elif [ "$v1_major" -gt "$v2_major" ]; then return 1; fi
 
-    # Compare minor version
-    if [ "$v1_minor" -lt "$v2_minor" ]; then
-        return 0
-    elif [ "$v1_minor" -gt "$v2_minor" ]; then
-        return 1
-    fi
+    if [ "$v1_minor" -lt "$v2_minor" ]; then return 0
+    elif [ "$v1_minor" -gt "$v2_minor" ]; then return 1; fi
 
-    # Compare patch version
-    if [ "$v1_patch" -le "$v2_patch" ]; then
-        return 0
-    else
-        return 1
-    fi
+    [ "$v1_patch" -le "$v2_patch" ]
 }
 
 # Check Docker daemon
@@ -336,10 +599,20 @@ fi
 
 # Convert to GB for display
 if [ "$MEMORY_MB" -gt 0 ]; then
-    MEMORY_GB=$((MEMORY_MB / 1024))
-    print_info "Docker memory allocation: ~${MEMORY_GB}GB"
+    MEMORY_GB=$(awk "BEGIN {printf \"%.1f\", $MEMORY_MB / 1024}")
+    if [ "$(awk "BEGIN {print ($MEMORY_MB >= 1024)}")" = "1" ]; then
+        MEMORY_DISPLAY="~${MEMORY_GB}GB"
+    else
+        MEMORY_DISPLAY="${MEMORY_MB}MB"
+    fi
+    if [[ "$OSTYPE" == "darwin"* ]]; then
+        print_info "Docker memory allocation: ${MEMORY_DISPLAY}"
+    else
+        print_info "System memory: ${MEMORY_DISPLAY} (Docker uses host memory directly)"
+    fi
 else
-    print_warning "Could not determine Docker memory allocation"
+    print_warning "Could not determine memory allocation"
+    MEMORY_DISPLAY="unknown"
     MEMORY_MB=0
 fi
 
@@ -358,7 +631,7 @@ RESOURCE_WARNING=false
 EXPECTED_RAM_MB=$((EXPECTED_DOCKER_RAM_GB * 1024))
 
 if [ "$MEMORY_MB" -gt 0 ] && [ "$MEMORY_MB" -lt "$EXPECTED_RAM_MB" ]; then
-    print_warning "Docker has less than ${EXPECTED_DOCKER_RAM_GB}GB RAM allocated (found: ~${MEMORY_GB}GB)"
+    print_warning "Less than ${EXPECTED_DOCKER_RAM_GB}GB RAM available (found: ${MEMORY_DISPLAY})"
     RESOURCE_WARNING=true
 fi
 
@@ -369,10 +642,10 @@ fi
 
 if [ "$RESOURCE_WARNING" = true ]; then
     echo ""
-    print_warning "Onyx recommends at least ${EXPECTED_DOCKER_RAM_GB}GB RAM and ${EXPECTED_DISK_GB}GB disk space for optimal performance."
-    echo ""
-    read -p "Do you want to continue anyway? (y/N): " -n 1 -r
+    print_warning "Onyx recommends at least ${EXPECTED_DOCKER_RAM_GB}GB RAM and ${EXPECTED_DISK_GB}GB disk space for optimal performance in standard mode."
+    print_warning "Lite mode requires less resources (1-4GB RAM, 8-16GB disk depending on usage), but does not include a vector database."
     echo ""
+    prompt_yn_or_default "Do you want to continue anyway? (Y/n): " "y"
     if [[ ! $REPLY =~ ^[Yy]$ ]]; then
         print_info "Installation cancelled. Please allocate more resources and try again."
         exit 1
@@ -385,117 +658,89 @@ print_step "Creating directory structure"
 if [ -d "${INSTALL_ROOT}" ]; then
     print_info "Directory structure already exists"
     print_success "Using existing ${INSTALL_ROOT} directory"
-else
-    mkdir -p "${INSTALL_ROOT}/deployment"
-    mkdir -p "${INSTALL_ROOT}/data/nginx/local"
-    print_success "Directory structure created"
 fi
+mkdir -p "${INSTALL_ROOT}/deployment"
+mkdir -p "${INSTALL_ROOT}/data/nginx/local"
+print_success "Directory structure created"
 
-# Download all required files
-print_step "Downloading Onyx configuration files"
-print_info "This step downloads all necessary configuration files from GitHub..."
-echo ""
-print_info "Downloading the following files:"
-echo "  • docker-compose.yml - Main Docker Compose configuration"
-echo "  • env.template - Environment variables template"
-echo "  • nginx/app.conf.template - Nginx web server configuration"
-echo "  • nginx/run-nginx.sh - Nginx startup script"
-echo "  • README.md - Documentation and setup instructions"
-echo ""
-
-# Download Docker Compose file
-COMPOSE_FILE="${INSTALL_ROOT}/deployment/docker-compose.yml"
-print_info "Downloading docker-compose.yml..."
-if curl -fsSL -o "$COMPOSE_FILE" "${GITHUB_RAW_URL}/docker-compose.yml" 2>/dev/null; then
-    print_success "Docker Compose file downloaded successfully"
-
-    # Check if Docker Compose version is older than 2.24.0 and show warning
-    # Skip check for dev builds (assume they're recent enough)
-    if [ "$COMPOSE_VERSION" != "dev" ] && version_compare "$COMPOSE_VERSION" "2.24.0"; then
-        print_warning "Docker Compose version $COMPOSE_VERSION is older than 2.24.0"
-        echo ""
-        print_warning "The docker-compose.yml file uses the newer env_file format that requires Docker Compose 2.24.0 or later."
-        echo ""
-        print_info "To use this configuration with your current Docker Compose version, you have two options:"
-        echo ""
-        echo "1. Upgrade Docker Compose to version 2.24.0 or later (recommended)"
-        echo "   Visit: https://docs.docker.com/compose/install/"
-        echo ""
-        echo "2. Manually replace all env_file sections in docker-compose.yml"
-        echo "   Change from:"
-        echo "     env_file:"
-        echo "       - path: .env"
-        echo "         required: false"
-        echo "   To:"
-        echo "     env_file: .env"
-        echo ""
-        print_warning "The installation will continue, but may fail if Docker Compose cannot parse the file."
-        echo ""
-        read -p "Do you want to continue anyway? (y/N): " -n 1 -r
-        echo ""
-        if [[ ! $REPLY =~ ^[Yy]$ ]]; then
-            print_info "Installation cancelled. Please upgrade Docker Compose or manually edit the docker-compose.yml file."
-            exit 1
-        fi
-        print_info "Proceeding with installation despite Docker Compose version compatibility issues..."
-    fi
-else
-    print_error "Failed to download Docker Compose file"
-    print_info "Please ensure you have internet connection and try again"
-    exit 1
-fi
-
-# Download env.template file
-ENV_TEMPLATE="${INSTALL_ROOT}/deployment/env.template"
-print_info "Downloading env.template..."
-if curl -fsSL -o "$ENV_TEMPLATE" "${GITHUB_RAW_URL}/env.template" 2>/dev/null; then
-    print_success "Environment template downloaded successfully"
-else
-    print_error "Failed to download env.template"
-    print_info "Please ensure you have internet connection and try again"
-    exit 1
-fi
-
-# Download nginx config files
+# Ensure all required configuration files are present
 NGINX_BASE_URL="https://raw.githubusercontent.com/onyx-dot-app/onyx/main/deployment/data/nginx"
 
-# Download app.conf.template
-NGINX_CONFIG="${INSTALL_ROOT}/data/nginx/app.conf.template"
-print_info "Downloading nginx configuration template..."
-if curl -fsSL -o "$NGINX_CONFIG" "$NGINX_BASE_URL/app.conf.template" 2>/dev/null; then
-    print_success "Nginx configuration template downloaded"
+if [[ "$USE_LOCAL_FILES" = true ]]; then
+    print_step "Verifying existing configuration files"
 else
-    print_error "Failed to download nginx configuration template"
-    print_info "Please ensure you have internet connection and try again"
-    exit 1
+    print_step "Downloading Onyx configuration files"
+    print_info "This step downloads all necessary configuration files from GitHub..."
 fi
 
-# Download run-nginx.sh script
-NGINX_RUN_SCRIPT="${INSTALL_ROOT}/data/nginx/run-nginx.sh"
-print_info "Downloading nginx startup script..."
-if curl -fsSL -o "$NGINX_RUN_SCRIPT" "$NGINX_BASE_URL/run-nginx.sh" 2>/dev/null; then
-    chmod +x "$NGINX_RUN_SCRIPT"
-    print_success "Nginx startup script downloaded and made executable"
-else
-    print_error "Failed to download nginx startup script"
-    print_info "Please ensure you have internet connection and try again"
-    exit 1
+ensure_file "${INSTALL_ROOT}/deployment/docker-compose.yml" \
+    "${GITHUB_RAW_URL}/docker-compose.yml" "docker-compose.yml" || exit 1
+
+# Check Docker Compose version compatibility after obtaining docker-compose.yml
+if [ "$COMPOSE_VERSION" != "dev" ] && version_compare "$COMPOSE_VERSION" "2.24.0"; then
+    print_warning "Docker Compose version $COMPOSE_VERSION is older than 2.24.0"
+    echo ""
+    print_warning "The docker-compose.yml file uses the newer env_file format that requires Docker Compose 2.24.0 or later."
+    echo ""
+    print_info "To use this configuration with your current Docker Compose version, you have two options:"
+    echo ""
+    echo "1. Upgrade Docker Compose to version 2.24.0 or later (recommended)"
+    echo "   Visit: https://docs.docker.com/compose/install/"
+    echo ""
+    echo "2. Manually replace all env_file sections in docker-compose.yml"
+    echo "   Change from:"
+    echo "     env_file:"
+    echo "       - path: .env"
+    echo "         required: false"
+    echo "   To:"
+    echo "     env_file: .env"
+    echo ""
+    print_warning "The installation will continue, but may fail if Docker Compose cannot parse the file."
+    echo ""
+    prompt_yn_or_default "Do you want to continue anyway? (Y/n): " "y"
+    if [[ ! $REPLY =~ ^[Yy]$ ]]; then
+        print_info "Installation cancelled. Please upgrade Docker Compose or manually edit the docker-compose.yml file."
+        exit 1
+    fi
+    print_info "Proceeding with installation despite Docker Compose version compatibility issues..."
 fi
 
-# Download README file
-README_FILE="${INSTALL_ROOT}/README.md"
-print_info "Downloading README.md..."
-if curl -fsSL -o "$README_FILE" "${GITHUB_RAW_URL}/README.md" 2>/dev/null; then
-    print_success "README.md downloaded successfully"
-else
-    print_error "Failed to download README.md"
-    print_info "Please ensure you have internet connection and try again"
-    exit 1
+# Handle lite overlay: ensure it if --lite, clean up stale copies otherwise
+if [[ "$LITE_MODE" = true ]]; then
+    ensure_file "${INSTALL_ROOT}/deployment/${LITE_COMPOSE_FILE}" \
+        "${GITHUB_RAW_URL}/${LITE_COMPOSE_FILE}" "${LITE_COMPOSE_FILE}" || exit 1
+elif [[ -f "${INSTALL_ROOT}/deployment/${LITE_COMPOSE_FILE}" ]]; then
+    if [[ -f "${INSTALL_ROOT}/deployment/.env" ]]; then
+        print_warning "Existing lite overlay found but --lite was not passed."
+        prompt_yn_or_default "Remove lite overlay and switch to standard mode? (y/N): " "n"
+        if [[ ! $REPLY =~ ^[Yy]$ ]]; then
+            print_info "Keeping existing lite overlay. Pass --lite to keep using lite mode."
+            LITE_MODE=true
+        else
+            rm -f "${INSTALL_ROOT}/deployment/${LITE_COMPOSE_FILE}"
+            print_info "Removed lite overlay (switching to standard mode)"
+        fi
+    else
+        rm -f "${INSTALL_ROOT}/deployment/${LITE_COMPOSE_FILE}"
+        print_info "Removed previous lite overlay (switching to standard mode)"
+    fi
 fi
 
-# Create empty local directory marker (if needed)
+ensure_file "${INSTALL_ROOT}/deployment/env.template" \
+    "${GITHUB_RAW_URL}/env.template" "env.template" || exit 1
+
+ensure_file "${INSTALL_ROOT}/data/nginx/app.conf.template" \
+    "$NGINX_BASE_URL/app.conf.template" "nginx/app.conf.template" || exit 1
+
+ensure_file "${INSTALL_ROOT}/data/nginx/run-nginx.sh" \
+    "$NGINX_BASE_URL/run-nginx.sh" "nginx/run-nginx.sh" || exit 1
+chmod +x "${INSTALL_ROOT}/data/nginx/run-nginx.sh"
+
+ensure_file "${INSTALL_ROOT}/README.md" \
+    "${GITHUB_RAW_URL}/README.md" "README.md" || exit 1
+
 touch "${INSTALL_ROOT}/data/nginx/local/.gitkeep"
-print_success "All configuration files downloaded successfully"
+print_success "All configuration files ready"
 
 # Set up deployment configuration
 print_step "Setting up deployment configs"
@@ -513,7 +758,7 @@ if [ -d "${INSTALL_ROOT}/deployment" ] && [ -f "${INSTALL_ROOT}/deployment/docke
 
     if [ -n "$COMPOSE_CMD" ]; then
         # Check if any containers are running
-        RUNNING_CONTAINERS=$(cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD -f docker-compose.yml ps -q 2>/dev/null | wc -l)
+        RUNNING_CONTAINERS=$(cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD $(compose_file_args true) ps -q 2>/dev/null | wc -l)
         if [ "$RUNNING_CONTAINERS" -gt 0 ]; then
             print_error "Onyx services are currently running!"
             echo ""
@@ -534,7 +779,7 @@ if [ -f "$ENV_FILE" ]; then
     echo "• Press Enter to restart with current configuration"
     echo "• Type 'update' to update to a newer version"
     echo ""
-    read -p "Choose an option [default: restart]: " -r
+    prompt_or_default "Choose an option [default: restart]: " ""
     echo ""
 
     if [ "$REPLY" = "update" ]; then
@@ -543,26 +788,30 @@ if [ -f "$ENV_FILE" ]; then
         echo "• Press Enter for latest (recommended)"
         echo "• Type a specific tag (e.g., v0.1.0)"
         echo ""
-        # If --include-craft was passed, default to craft-latest
         if [ "$INCLUDE_CRAFT" = true ]; then
-            read -p "Enter tag [default: craft-latest]: " -r VERSION
+            prompt_or_default "Enter tag [default: craft-latest]: " "craft-latest"
+            VERSION="$REPLY"
         else
-            read -p "Enter tag [default: latest]: " -r VERSION
+            prompt_or_default "Enter tag [default: latest]: " "latest"
+            VERSION="$REPLY"
         fi
         echo ""
 
-        if [ -z "$VERSION" ]; then
-            if [ "$INCLUDE_CRAFT" = true ]; then
-                VERSION="craft-latest"
-                print_info "Selected: craft-latest (Craft enabled)"
-            else
-                VERSION="latest"
-                print_info "Selected: Latest version"
-            fi
+        if [ "$INCLUDE_CRAFT" = true ] && [ "$VERSION" = "craft-latest" ]; then
+            print_info "Selected: craft-latest (Craft enabled)"
+        elif [ "$VERSION" = "latest" ]; then
+            print_info "Selected: Latest version"
         else
             print_info "Selected: $VERSION"
         fi
 
+        # Reject craft image tags when running in lite mode
+        if [[ "$LITE_MODE" = true ]] && [[ "${VERSION:-}" == craft-* ]]; then
+            print_error "Cannot use a craft image tag (${VERSION}) with --lite."
+            print_info "Craft requires services (Vespa, Redis, background workers) that lite mode disables."
+            exit 1
+        fi
+
         # Update .env file with new version
         print_info "Updating configuration for version $VERSION..."
         if grep -q "^IMAGE_TAG=" "$ENV_FILE"; then
@@ -581,13 +830,67 @@ if [ -f "$ENV_FILE" ]; then
         fi
         print_success "Configuration updated for upgrade"
     else
+        # Reject restarting a craft deployment in lite mode
+        EXISTING_TAG=$(grep "^IMAGE_TAG=" "$ENV_FILE" | head -1 | cut -d'=' -f2 | tr -d ' "'"'"'')
+        if [[ "$LITE_MODE" = true ]] && [[ "${EXISTING_TAG:-}" == craft-* ]]; then
+            print_error "Cannot restart a craft deployment (${EXISTING_TAG}) with --lite."
+            print_info "Craft requires services (Vespa, Redis, background workers) that lite mode disables."
+            exit 1
+        fi
+
         print_info "Keeping existing configuration..."
         print_success "Will restart with current settings"
     fi
+
+    # Ensure COMPOSE_PROFILES is cleared when running in lite mode on an
+    # existing .env (the template ships with s3-filestore enabled).
+    if [[ "$LITE_MODE" = true ]] && grep -q "^COMPOSE_PROFILES=.*s3-filestore" "$ENV_FILE" 2>/dev/null; then
+        sed -i.bak 's/^COMPOSE_PROFILES=.*/COMPOSE_PROFILES=/' "$ENV_FILE" 2>/dev/null || true
+        print_success "Cleared COMPOSE_PROFILES for lite mode"
+    fi
 else
     print_info "No existing .env file found. Setting up new deployment..."
     echo ""
 
+    # Ask for deployment mode (standard vs lite) unless already set via --lite flag
+    if [[ "$LITE_MODE" = false ]]; then
+        print_info "Which deployment mode would you like?"
+        echo ""
+        echo "  1) Standard  - Full deployment with search, connectors, and RAG"
+        echo "  2) Lite      - Minimal deployment (no Vespa, Redis, or model servers)"
+        echo "                  LLM chat, tools, file uploads, and Projects still work"
+        echo ""
+        prompt_or_default "Choose a mode (1 or 2) [default: 1]: " "1"
+        echo ""
+
+        case "$REPLY" in
+            2)
+                LITE_MODE=true
+                print_info "Selected: Lite mode"
+                ensure_file "${INSTALL_ROOT}/deployment/${LITE_COMPOSE_FILE}" \
+                    "${GITHUB_RAW_URL}/${LITE_COMPOSE_FILE}" "${LITE_COMPOSE_FILE}" || exit 1
+                ;;
+            *)
+                print_info "Selected: Standard mode"
+                ;;
+        esac
+    else
+        print_info "Deployment mode: Lite (set via --lite flag)"
+    fi
+
+    # Validate lite + craft combination (could now be set interactively)
+    if [[ "$LITE_MODE" = true ]] && [[ "$INCLUDE_CRAFT" = true ]]; then
+        print_error "--include-craft cannot be used with Lite mode."
+        print_info "Craft requires services (Vespa, Redis, background workers) that lite mode disables."
+        exit 1
+    fi
+
+    # Adjust resource expectations for lite mode
+    if [[ "$LITE_MODE" = true ]]; then
+        EXPECTED_DOCKER_RAM_GB=4
+        EXPECTED_DISK_GB=16
+    fi
+
     # Ask for version
     print_info "Which tag would you like to deploy?"
     echo ""
@@ -595,23 +898,21 @@ else
         echo "• Press Enter for craft-latest (recommended for Craft)"
         echo "• Type a specific tag (e.g., craft-v1.0.0)"
         echo ""
-        read -p "Enter tag [default: craft-latest]: " -r VERSION
+        prompt_or_default "Enter tag [default: craft-latest]: " "craft-latest"
+        VERSION="$REPLY"
     else
         echo "• Press Enter for latest (recommended)"
         echo "• Type a specific tag (e.g., v0.1.0)"
         echo ""
-        read -p "Enter tag [default: latest]: " -r VERSION
+        prompt_or_default "Enter tag [default: latest]: " "latest"
+        VERSION="$REPLY"
     fi
     echo ""
 
-    if [ -z "$VERSION" ]; then
-        if [ "$INCLUDE_CRAFT" = true ]; then
-            VERSION="craft-latest"
-            print_info "Selected: craft-latest (Craft enabled)"
-        else
-            VERSION="latest"
-            print_info "Selected: Latest tag"
-        fi
+    if [ "$INCLUDE_CRAFT" = true ] && [ "$VERSION" = "craft-latest" ]; then
+        print_info "Selected: craft-latest (Craft enabled)"
+    elif [ "$VERSION" = "latest" ]; then
+        print_info "Selected: Latest tag"
     else
         print_info "Selected: $VERSION"
     fi
@@ -645,6 +946,13 @@ else
     # Use basic auth by default
     AUTH_SCHEMA="basic"
 
+    # Reject craft image tags when running in lite mode (must check before writing .env)
+    if [[ "$LITE_MODE" = true ]] && [[ "${VERSION:-}" == craft-* ]]; then
+        print_error "Cannot use a craft image tag (${VERSION}) with --lite."
+        print_info "Craft requires services (Vespa, Redis, background workers) that lite mode disables."
+        exit 1
+    fi
+
     # Create .env file from template
     print_info "Creating .env file with your selections..."
     cp "$ENV_TEMPLATE" "$ENV_FILE"
@@ -654,6 +962,13 @@ else
     sed -i.bak "s/^IMAGE_TAG=.*/IMAGE_TAG=$VERSION/" "$ENV_FILE"
     print_success "IMAGE_TAG set to $VERSION"
 
+    # In lite mode, clear COMPOSE_PROFILES so profiled services (MinIO, etc.)
+    # stay disabled — the template ships with s3-filestore enabled by default.
+    if [[ "$LITE_MODE" = true ]]; then
+        sed -i.bak 's/^COMPOSE_PROFILES=.*/COMPOSE_PROFILES=/' "$ENV_FILE" 2>/dev/null || true
+        print_success "Cleared COMPOSE_PROFILES for lite mode"
+    fi
+
     # Configure basic authentication (default)
     sed -i.bak 's/^AUTH_TYPE=.*/AUTH_TYPE=basic/' "$ENV_FILE" 2>/dev/null || true
     print_success "Basic authentication enabled in configuration"
@@ -774,7 +1089,7 @@ print_step "Pulling Docker images"
 print_info "This may take several minutes depending on your internet connection..."
 echo ""
 print_info "Downloading Docker images (this may take a while)..."
-(cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD -f docker-compose.yml pull --quiet)
+(cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD $(compose_file_args) pull --quiet)
 if [ $? -eq 0 ]; then
     print_success "Docker images downloaded successfully"
 else
@@ -788,9 +1103,9 @@ print_info "Launching containers..."
 echo ""
 if [ "$USE_LATEST" = true ]; then
     print_info "Force pulling latest images and recreating containers..."
-    (cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD -f docker-compose.yml up -d --pull always --force-recreate)
+    (cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD $(compose_file_args) up -d --pull always --force-recreate)
 else
-    (cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD -f docker-compose.yml up -d)
+    (cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD $(compose_file_args) up -d)
 fi
 if [ $? -ne 0 ]; then
     print_error "Failed to start Onyx services"
@@ -812,7 +1127,7 @@ echo ""
 # Check for restart loops
 print_info "Checking container health status..."
 RESTART_ISSUES=false
-CONTAINERS=$(cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD -f docker-compose.yml ps -q 2>/dev/null)
+CONTAINERS=$(cd "${INSTALL_ROOT}/deployment" && $COMPOSE_CMD $(compose_file_args) ps -q 2>/dev/null)
 
 for CONTAINER in $CONTAINERS; do
     PROJECT_NAME="$(basename "$INSTALL_ROOT")_deployment_"
@@ -841,7 +1156,7 @@ if [ "$RESTART_ISSUES" = true ]; then
     print_error "Some containers are experiencing issues!"
     echo ""
     print_info "Please check the logs for more information:"
-    echo "  (cd \"${INSTALL_ROOT}/deployment\" && $COMPOSE_CMD -f docker-compose.yml logs)"
+    echo "  (cd \"${INSTALL_ROOT}/deployment\" && $COMPOSE_CMD $(compose_file_args) logs)"
 
     echo ""
     print_info "If the issue persists, please contact: founders@onyx.app"
@@ -860,8 +1175,12 @@ check_onyx_health() {
     echo ""
 
     while [ $attempt -le $max_attempts ]; do
-        # Check for successful HTTP responses (200, 301, 302, etc.)
-        local http_code=$(curl -s -o /dev/null -w "%{http_code}" "http://localhost:$port")
+        local http_code=""
+        if [[ "$DOWNLOADER" == "curl" ]]; then
+            http_code=$(curl -s -o /dev/null -w "%{http_code}" "http://localhost:$port" 2>/dev/null || echo "000")
+        else
+            http_code=$(wget -q --spider -S "http://localhost:$port" 2>&1 | grep "HTTP/" | tail -1 | awk '{print $2}' || echo "000")
+        fi
         if echo "$http_code" | grep -qE "^(200|301|302|303|307|308)$"; then
             return 0
         fi
@@ -917,6 +1236,18 @@ print_info "If authentication is enabled, you can create your admin account here
 echo "   • Visit http://localhost:${HOST_PORT}/auth/signup to create your admin account"
 echo "   • The first user created will automatically have admin privileges"
 echo ""
+if [[ "$LITE_MODE" = true ]]; then
+    echo ""
+    print_info "Running in Lite mode — the following services are NOT started:"
+    echo "  • Vespa (vector database)"
+    echo "  • Redis (cache)"
+    echo "  • Model servers (embedding/inference)"
+    echo "  • Background workers (Celery)"
+    echo ""
+    print_info "Connectors and RAG search are disabled. LLM chat, tools, user file"
+    print_info "uploads, Projects, Agent knowledge, and code interpreter still work."
+fi
+echo ""
 print_info "Refer to the README in the ${INSTALL_ROOT} directory for more information."
 echo ""
 print_info "For help or issues, contact: founders@onyx.app"

pyproject.toml

@@ -91,7 +91,7 @@ backend = [
     "python-gitlab==5.6.0",
     "python-pptx==0.6.23",
     "pypandoc_binary==1.16.2",
-    "pypdf==6.8.0",
+    "pypdf==6.7.5",
     "pytest-mock==3.12.0",
     "pytest-playwright==0.7.0",
     "python-docx==1.1.2",
@@ -143,7 +143,7 @@ dev = [
     "matplotlib==3.10.8",
     "mypy-extensions==1.0.0",
     "mypy==1.13.0",
-    "onyx-devtools==0.7.0",
+    "onyx-devtools==0.6.3",
     "openapi-generator-cli==7.17.0",
     "pandas-stubs~=2.3.3",
     "pre-commit==3.2.2",
@@ -153,7 +153,7 @@ dev = [
     "pytest-repeat==0.9.4",
     "pytest-xdist==3.8.0",
     "pytest==8.3.5",
-    "release-tag==0.5.2",
+    "release-tag==0.4.3",
     "reorder-python-imports-black==3.14.0",
     "ruff==0.12.0",
     "types-beautifulsoup4==4.12.0.3",

tools/ods/README.md

@@ -25,9 +25,6 @@ Some commands require external tools to be installed and configured:
 - **Docker** - Required for `compose`, `logs`, and `pull` commands
   - Install from [docker.com](https://docs.docker.com/get-docker/)
 
-- **uv** - Required for `backend` commands
-  - Install from [docs.astral.sh/uv](https://docs.astral.sh/uv/)
-
 - **GitHub CLI** (`gh`) - Required for `run-ci` and `cherry-pick` commands
   - Install from [cli.github.com](https://cli.github.com/)
   - Authenticate with `gh auth login`
@@ -173,53 +170,6 @@ ods pull
 ods pull --tag edge
 ```
 
-### `backend` - Run Backend Services
-
-Run backend services (API server, model server) with environment loaded from
-`.vscode/.env`. On first run, copies `.vscode/env_template.txt` to `.vscode/.env`
-if the `.env` file does not already exist.
-
-Enterprise Edition features are enabled by default with license enforcement
-disabled, matching the `compose` command behavior.
-
-```shell
-ods backend <subcommand>
-```
-
-**Subcommands:**
-
-- `api` - Start the FastAPI backend server (`uvicorn onyx.main:app --reload`)
-- `model_server` - Start the model server (`uvicorn model_server.main:app --reload`)
-
-**Flags:**
-
-| Flag | Default | Description |
-|------|---------|-------------|
-| `--no-ee` | `false` | Disable Enterprise Edition features (enabled by default) |
-| `--port` | `8080` (api) / `9000` (model_server) | Port to listen on |
-
-Shell environment takes precedence over `.env` file values, so inline overrides
-work as expected (e.g. `S3_ENDPOINT_URL=foo ods backend api`).
-
-**Examples:**
-
-```shell
-# Start the API server
-ods backend api
-
-# Start the API server on a custom port
-ods backend api --port 9090
-
-# Start without Enterprise Edition
-ods backend api --no-ee
-
-# Start the model server
-ods backend model_server
-
-# Start the model server on a custom port
-ods backend model_server --port 9001
-```
-
 ### `web` - Run Frontend Scripts
 
 Run npm scripts from `web/package.json` without manually changing directories.

tools/ods/cmd/backend.go

@@ -1,242 +0,0 @@
-package cmd
-
-import (
-	"bufio"
-	"errors"
-	"fmt"
-	"os"
-	"os/exec"
-	"path/filepath"
-	"strings"
-
-	log "github.com/sirupsen/logrus"
-	"github.com/spf13/cobra"
-
-	"github.com/onyx-dot-app/onyx/tools/ods/internal/paths"
-)
-
-// NewBackendCommand creates the parent "backend" command with subcommands for
-// running backend services.
-// BackendOptions holds options shared across backend subcommands.
-type BackendOptions struct {
-	NoEE bool
-}
-
-func NewBackendCommand() *cobra.Command {
-	opts := &BackendOptions{}
-
-	cmd := &cobra.Command{
-		Use:   "backend",
-		Short: "Run backend services (api, model_server)",
-		Long: `Run backend services with environment from .vscode/.env.
-
-On first run, copies .vscode/env_template.txt to .vscode/.env if the
-.env file does not already exist.
-
-Enterprise Edition features are enabled by default for development,
-with license enforcement disabled.
-
-Available subcommands:
-  api            Start the FastAPI backend server
-  model_server   Start the model server`,
-	}
-
-	cmd.PersistentFlags().BoolVar(&opts.NoEE, "no-ee", false, "Disable Enterprise Edition features (enabled by default)")
-
-	cmd.AddCommand(newBackendAPICommand(opts))
-	cmd.AddCommand(newBackendModelServerCommand(opts))
-
-	return cmd
-}
-
-func newBackendAPICommand(opts *BackendOptions) *cobra.Command {
-	var port string
-
-	cmd := &cobra.Command{
-		Use:   "api",
-		Short: "Start the backend API server (uvicorn with hot-reload)",
-		Long: `Start the backend API server using uvicorn with hot-reload.
-
-Examples:
-  ods backend api
-  ods backend api --port 9090
-  ods backend api --no-ee`,
-		Run: func(cmd *cobra.Command, args []string) {
-			runBackendService("api", "onyx.main:app", port, opts)
-		},
-	}
-
-	cmd.Flags().StringVar(&port, "port", "8080", "Port to listen on")
-
-	return cmd
-}
-
-func newBackendModelServerCommand(opts *BackendOptions) *cobra.Command {
-	var port string
-
-	cmd := &cobra.Command{
-		Use:   "model_server",
-		Short: "Start the model server (uvicorn with hot-reload)",
-		Long: `Start the model server using uvicorn with hot-reload.
-
-Examples:
-  ods backend model_server
-  ods backend model_server --port 9001`,
-		Run: func(cmd *cobra.Command, args []string) {
-			runBackendService("model_server", "model_server.main:app", port, opts)
-		},
-	}
-
-	cmd.Flags().StringVar(&port, "port", "9000", "Port to listen on")
-
-	return cmd
-}
-
-func runBackendService(name, module, port string, opts *BackendOptions) {
-	root, err := paths.GitRoot()
-	if err != nil {
-		log.Fatalf("Failed to find git root: %v", err)
-	}
-
-	envFile := ensureBackendEnvFile(root)
-	fileVars := loadBackendEnvFile(envFile)
-
-	eeDefaults := eeEnvDefaults(opts.NoEE)
-	fileVars = append(fileVars, eeDefaults...)
-
-	backendDir := filepath.Join(root, "backend")
-
-	uvicornArgs := []string{
-		"run", "uvicorn", module,
-		"--reload",
-		"--port", port,
-	}
-	log.Infof("Starting %s on port %s...", name, port)
-	if !opts.NoEE {
-		log.Info("Enterprise Edition enabled (use --no-ee to disable)")
-	}
-	log.Debugf("Running in %s: uv %v", backendDir, uvicornArgs)
-
-	mergedEnv := mergeEnv(os.Environ(), fileVars)
-	log.Debugf("Applied %d env vars from %s (shell takes precedence)", len(fileVars), envFile)
-
-	svcCmd := exec.Command("uv", uvicornArgs...)
-	svcCmd.Dir = backendDir
-	svcCmd.Stdout = os.Stdout
-	svcCmd.Stderr = os.Stderr
-	svcCmd.Stdin = os.Stdin
-	svcCmd.Env = mergedEnv
-
-	if err := svcCmd.Run(); err != nil {
-		var exitErr *exec.ExitError
-		if errors.As(err, &exitErr) {
-			if code := exitErr.ExitCode(); code != -1 {
-				os.Exit(code)
-			}
-		}
-		log.Fatalf("Failed to run %s: %v", name, err)
-	}
-}
-
-// eeEnvDefaults returns env entries for EE and license enforcement settings.
-// These are appended to the file vars so they act as defaults — shell env
-// and .env file values still take precedence via mergeEnv.
-func eeEnvDefaults(noEE bool) []string {
-	if noEE {
-		return []string{
-			"ENABLE_PAID_ENTERPRISE_EDITION_FEATURES=false",
-		}
-	}
-	return []string{
-		"ENABLE_PAID_ENTERPRISE_EDITION_FEATURES=true",
-		"LICENSE_ENFORCEMENT_ENABLED=false",
-	}
-}
-
-// ensureBackendEnvFile copies env_template.txt to .env if .env doesn't exist.
-func ensureBackendEnvFile(root string) string {
-	vscodeDir := filepath.Join(root, ".vscode")
-	envFile := filepath.Join(vscodeDir, ".env")
-	templateFile := filepath.Join(vscodeDir, "env_template.txt")
-
-	if _, err := os.Stat(envFile); err != nil {
-		if !errors.Is(err, os.ErrNotExist) {
-			log.Fatalf("Failed to stat env file %s: %v", envFile, err)
-		}
-	} else {
-		log.Debugf("Using existing env file: %s", envFile)
-		return envFile
-	}
-
-	templateData, err := os.ReadFile(templateFile)
-	if err != nil {
-		log.Fatalf("Failed to read env template %s: %v", templateFile, err)
-	}
-
-	if err := os.MkdirAll(vscodeDir, 0755); err != nil {
-		log.Fatalf("Failed to create .vscode directory: %v", err)
-	}
-
-	if err := os.WriteFile(envFile, templateData, 0644); err != nil {
-		log.Fatalf("Failed to write env file %s: %v", envFile, err)
-	}
-
-	log.Infof("Created %s from template (review and fill in <REPLACE THIS> values)", envFile)
-	return envFile
-}
-
-// mergeEnv combines shell environment with file-based defaults. Shell values
-// take precedence — file entries are only added for keys not already present.
-func mergeEnv(shellEnv, fileVars []string) []string {
-	existing := make(map[string]bool, len(shellEnv))
-	for _, entry := range shellEnv {
-		if idx := strings.Index(entry, "="); idx > 0 {
-			existing[entry[:idx]] = true
-		}
-	}
-
-	merged := make([]string, len(shellEnv))
-	copy(merged, shellEnv)
-	for _, entry := range fileVars {
-		if idx := strings.Index(entry, "="); idx > 0 {
-			key := entry[:idx]
-			if !existing[key] {
-				merged = append(merged, entry)
-			} else {
-				log.Debugf("Env var %s already set in shell, skipping .env value", key)
-			}
-		}
-	}
-	return merged
-}
-
-// loadBackendEnvFile parses a .env file into KEY=VALUE entries suitable for
-// appending to os.Environ(). Blank lines and comments are skipped.
-func loadBackendEnvFile(path string) []string {
-	f, err := os.Open(path)
-	if err != nil {
-		log.Fatalf("Failed to open env file %s: %v", path, err)
-	}
-	defer func() { _ = f.Close() }()
-
-	var envVars []string
-	scanner := bufio.NewScanner(f)
-	for scanner.Scan() {
-		line := strings.TrimSpace(scanner.Text())
-		if line == "" || strings.HasPrefix(line, "#") {
-			continue
-		}
-		if idx := strings.Index(line, "="); idx > 0 {
-			key := strings.TrimSpace(line[:idx])
-			value := strings.TrimSpace(line[idx+1:])
-			value = strings.Trim(value, `"'`)
-			envVars = append(envVars, fmt.Sprintf("%s=%s", key, value))
-		}
-	}
-
-	if err := scanner.Err(); err != nil {
-		log.Fatalf("Failed to read env file %s: %v", path, err)
-	}
-
-	return envVars
-}

tools/ods/cmd/print_latest.go

@@ -1,36 +0,0 @@
-package cmd
-
-import (
-	"fmt"
-
-	"github.com/jmelahman/tag/git"
-	"github.com/spf13/cobra"
-)
-
-// NewLatestStableTagCommand creates the latest-stable-tag command.
-func NewLatestStableTagCommand() *cobra.Command {
-	cmd := &cobra.Command{
-		Use:   "latest-stable-tag",
-		Short: "Print the git tag that should receive the 'latest' Docker tag",
-		Long: `Print the highest stable (non-pre-release) semver tag in the repository.
-
-This is used during deployment to decide whether a given tag should
-receive the "latest" tag on Docker Hub. Only the highest vX.Y.Z tag
-qualifies. Tags with pre-release suffixes (e.g. v1.2.3-beta,
-v1.2.3-cloud.1) are excluded.`,
-		Args: cobra.NoArgs,
-		RunE: func(c *cobra.Command, _ []string) error {
-			tag, err := git.GetLatestStableSemverTag("")
-			if err != nil {
-				return fmt.Errorf("get latest stable semver tag: %w", err)
-			}
-			if tag == "" {
-				return fmt.Errorf("no stable semver tag found in repository")
-			}
-			fmt.Println(tag)
-			return nil
-		},
-	}
-
-	return cmd
-}

tools/ods/cmd/root.go

@@ -41,7 +41,6 @@ func NewRootCommand() *cobra.Command {
 	cmd.PersistentFlags().BoolVar(&opts.Debug, "debug", false, "run in debug mode")
 
 	// Add subcommands
-	cmd.AddCommand(NewBackendCommand())
 	cmd.AddCommand(NewCheckLazyImportsCommand())
 	cmd.AddCommand(NewCherryPickCommand())
 	cmd.AddCommand(NewDBCommand())
@@ -53,7 +52,6 @@ func NewRootCommand() *cobra.Command {
 	cmd.AddCommand(NewScreenshotDiffCommand())
 	cmd.AddCommand(NewDesktopCommand())
 	cmd.AddCommand(NewWebCommand())
-	cmd.AddCommand(NewLatestStableTagCommand())
 	cmd.AddCommand(NewWhoisCommand())
 
 	return cmd

tools/ods/go.mod

@@ -3,13 +3,12 @@ module github.com/onyx-dot-app/onyx/tools/ods
 go 1.26.0
 
 require (
-	github.com/jmelahman/tag v0.5.2
 	github.com/sirupsen/logrus v1.9.3
-	github.com/spf13/cobra v1.10.2
-	github.com/spf13/pflag v1.0.10
+	github.com/spf13/cobra v1.10.1
+	github.com/spf13/pflag v1.0.9
 )
 
 require (
 	github.com/inconshreveable/mousetrap v1.1.0 // indirect
-	golang.org/x/sys v0.39.0 // indirect
+	golang.org/x/sys v0.0.0-20220715151400-c0bba94af5f8 // indirect
 )

tools/ods/go.sum

@@ -4,26 +4,20 @@ github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/inconshreveable/mousetrap v1.1.0 h1:wN+x4NVGpMsO7ErUn/mUI3vEoE6Jt13X2s0bqwp9tc8=
 github.com/inconshreveable/mousetrap v1.1.0/go.mod h1:vpF70FUmC8bwa3OWnCshd2FqLfsEA9PFc4w1p2J65bw=
-github.com/jmelahman/tag v0.5.2 h1:g6A/aHehu5tkA31mPoDsXBNr1FigZ9A82Y8WVgb/WsM=
-github.com/jmelahman/tag v0.5.2/go.mod h1:qmuqk19B1BKkpcg3kn7l/Eey+UqucLxgOWkteUGiG4Q=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/russross/blackfriday/v2 v2.1.0/go.mod h1:+Rmxgy9KzJVeS9/2gXHxylqXiyQDYRxCVz55jmeOWTM=
 github.com/sirupsen/logrus v1.9.3 h1:dueUQJ1C2q9oE3F7wvmSGAaVtTmUizReu6fjN8uqzbQ=
 github.com/sirupsen/logrus v1.9.3/go.mod h1:naHLuLoDiP4jHNo9R0sCBMtWGeIprob74mVsIT4qYEQ=
-github.com/spf13/cobra v1.10.2 h1:DMTTonx5m65Ic0GOoRY2c16WCbHxOOw6xxezuLaBpcU=
-github.com/spf13/cobra v1.10.2/go.mod h1:7C1pvHqHw5A4vrJfjNwvOdzYu0Gml16OCs2GRiTUUS4=
+github.com/spf13/cobra v1.10.1 h1:lJeBwCfmrnXthfAupyUTzJ/J4Nc1RsHC/mSRU2dll/s=
+github.com/spf13/cobra v1.10.1/go.mod h1:7SmJGaTHFVBY0jW4NXGluQoLvhqFQM+6XSKD+P4XaB0=
+github.com/spf13/pflag v1.0.9 h1:9exaQaMOCwffKiiiYk6/BndUBv+iRViNW+4lEMi0PvY=
 github.com/spf13/pflag v1.0.9/go.mod h1:McXfInJRrz4CZXVZOBLb0bTZqETkiAhM9Iw0y3An2Bg=
-github.com/spf13/pflag v1.0.10 h1:4EBh2KAYBwaONj6b2Ye1GiHfwjqyROoF4RwYO+vPwFk=
-github.com/spf13/pflag v1.0.10/go.mod h1:McXfInJRrz4CZXVZOBLb0bTZqETkiAhM9Iw0y3An2Bg=
 github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
+github.com/stretchr/testify v1.7.0 h1:nwc3DEeHmmLAfoZucVR881uASk0Mfjw8xYJ99tb5CcY=
 github.com/stretchr/testify v1.7.0/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
-github.com/stretchr/testify v1.11.1 h1:7s2iGBzp5EwR7/aIZr8ao5+dra3wiQyKjjFuvgVKu7U=
-github.com/stretchr/testify v1.11.1/go.mod h1:wZwfW3scLgRK+23gO65QZefKpKQRnfz6sD981Nm4B6U=
-go.yaml.in/yaml/v3 v3.0.4/go.mod h1:DhzuOOF2ATzADvBadXxruRBLzYTpT36CKvDb3+aBEFg=
+golang.org/x/sys v0.0.0-20220715151400-c0bba94af5f8 h1:0A+M6Uqn+Eje4kHMK80dtF3JCXC4ykBgQG4Fe06QRhQ=
 golang.org/x/sys v0.0.0-20220715151400-c0bba94af5f8/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.39.0 h1:CvCKL8MeisomCi6qNZ+wbb0DN9E5AATixKsvNtMoMFk=
-golang.org/x/sys v0.39.0/go.mod h1:OgkHotnGiDImocRcuBABYBEXf8A9a87e/uXjp9XT3ks=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
 gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=

uv.lock

@@ -4443,7 +4443,7 @@ requires-dist = [
     { name = "numpy", marker = "extra == 'model-server'", specifier = "==2.4.1" },
     { name = "oauthlib", marker = "extra == 'backend'", specifier = "==3.2.2" },
     { name = "office365-rest-python-client", marker = "extra == 'backend'", specifier = "==2.6.2" },
-    { name = "onyx-devtools", marker = "extra == 'dev'", specifier = "==0.7.0" },
+    { name = "onyx-devtools", marker = "extra == 'dev'", specifier = "==0.6.3" },
     { name = "openai", specifier = "==2.14.0" },
     { name = "openapi-generator-cli", marker = "extra == 'dev'", specifier = "==7.17.0" },
     { name = "openinference-instrumentation", marker = "extra == 'backend'", specifier = "==0.1.42" },
@@ -4466,7 +4466,7 @@ requires-dist = [
     { name = "pygithub", marker = "extra == 'backend'", specifier = "==2.5.0" },
     { name = "pympler", marker = "extra == 'backend'", specifier = "==1.1" },
     { name = "pypandoc-binary", marker = "extra == 'backend'", specifier = "==1.16.2" },
-    { name = "pypdf", marker = "extra == 'backend'", specifier = "==6.8.0" },
+    { name = "pypdf", marker = "extra == 'backend'", specifier = "==6.7.5" },
     { name = "pytest", marker = "extra == 'dev'", specifier = "==8.3.5" },
     { name = "pytest-alembic", marker = "extra == 'dev'", specifier = "==0.12.1" },
     { name = "pytest-asyncio", marker = "extra == 'dev'", specifier = "==1.3.0" },
@@ -4485,7 +4485,7 @@ requires-dist = [
     { name = "pywikibot", marker = "extra == 'backend'", specifier = "==9.0.0" },
     { name = "rapidfuzz", marker = "extra == 'backend'", specifier = "==3.13.0" },
     { name = "redis", marker = "extra == 'backend'", specifier = "==5.0.8" },
-    { name = "release-tag", marker = "extra == 'dev'", specifier = "==0.5.2" },
+    { name = "release-tag", marker = "extra == 'dev'", specifier = "==0.4.3" },
     { name = "reorder-python-imports-black", marker = "extra == 'dev'", specifier = "==3.14.0" },
     { name = "requests", marker = "extra == 'backend'", specifier = "==2.32.5" },
     { name = "requests-oauthlib", marker = "extra == 'backend'", specifier = "==1.3.1" },
@@ -4548,19 +4548,20 @@ requires-dist = [{ name = "onyx", extras = ["backend", "dev", "ee"], editable =
 
 [[package]]
 name = "onyx-devtools"
-version = "0.7.0"
+version = "0.6.3"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "fastapi" },
     { name = "openapi-generator-cli" },
 ]
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/22/9e/6957b11555da57d9e97092f4cd8ac09a86666264b0c9491838f4b27db5dc/onyx_devtools-0.7.0-py3-none-macosx_10_12_x86_64.whl", hash = "sha256:ad962a168d46ea11dcde9fa3b37e4f12ec520b4a4cb4d49d8732de110d46c4b6", size = 3998057, upload-time = "2026-03-12T03:09:11.585Z" },
-    { url = "https://files.pythonhosted.org/packages/cd/90/c72f3d06ba677012d77c77de36195b6a32a15c755c79ba0282be74e3c366/onyx_devtools-0.7.0-py3-none-macosx_11_0_arm64.whl", hash = "sha256:e46d252e2b048ff053b03519c3a875998780738d7c334eaa1c9a32ff445e3e1a", size = 3687753, upload-time = "2026-03-12T03:09:11.742Z" },
-    { url = "https://files.pythonhosted.org/packages/10/42/4e9fe36eccf9f76d67ba8f4ff6539196a09cd60351fb63f5865e1544cbfa/onyx_devtools-0.7.0-py3-none-manylinux_2_17_aarch64.whl", hash = "sha256:f280bc9320e1cc310e7d753a371009bfaab02cc0e0cfd78559663b15655b5a50", size = 3560144, upload-time = "2026-03-12T03:12:24.02Z" },
-    { url = "https://files.pythonhosted.org/packages/76/40/36dc12d99760b358c7f39b27361cb18fa9681ffe194107f982d0e1a74016/onyx_devtools-0.7.0-py3-none-manylinux_2_17_x86_64.whl", hash = "sha256:e31df751c7540ae7e70a7fe8e1153c79c31c2254af6aa4c72c0dd54fa381d2ab", size = 3964387, upload-time = "2026-03-12T03:09:11.356Z" },
-    { url = "https://files.pythonhosted.org/packages/34/18/74744230c3820a5a7687335507ca5f1dbebab2c5325805041c1cd5703e6a/onyx_devtools-0.7.0-py3-none-win_amd64.whl", hash = "sha256:541bfd347c2d5b11e7f63ab5001d2594df91d215ad9d07b1562f5e715700f7e6", size = 4068030, upload-time = "2026-03-12T03:09:12.98Z" },
-    { url = "https://files.pythonhosted.org/packages/8c/78/1320436607d3ffcb321ba7b064556c020ea15843a7e7d903fbb7529a71f5/onyx_devtools-0.7.0-py3-none-win_arm64.whl", hash = "sha256:83016330a9d39712431916cc25b2fb2cfcaa0112a55cc4f919d545da3a8974f9", size = 3626409, upload-time = "2026-03-12T03:09:10.222Z" },
+    { url = "https://files.pythonhosted.org/packages/84/e2/e7619722c3ccd18eb38100f776fb3dd6b4ae0fbbee09fca5af7c69a279b5/onyx_devtools-0.6.3-py3-none-any.whl", hash = "sha256:d3a5422945d9da12cafc185f64b39f6e727ee4cc92b37427deb7a38f9aad4966", size = 3945381, upload-time = "2026-03-05T20:39:25.896Z" },
+    { url = "https://files.pythonhosted.org/packages/f2/09/513d2dabedc1e54ad4376830fc9b34a3d9c164bdbcdedfcdbb8b8154dc5a/onyx_devtools-0.6.3-py3-none-macosx_10_12_x86_64.whl", hash = "sha256:efe300e9f3a2e7ae75f88a4f9e0a5c4c471478296cb1615b6a1f03d247582e13", size = 3978761, upload-time = "2026-03-05T20:39:28.822Z" },
+    { url = "https://files.pythonhosted.org/packages/39/41/e757602a0de032d74ed01c7ee57f30e57728fb9cd4f922f50d2affda3889/onyx_devtools-0.6.3-py3-none-macosx_11_0_arm64.whl", hash = "sha256:594066eed3f917cfab5a8c7eac3d4a210df30259f2049f664787749709345e19", size = 3665378, upload-time = "2026-03-05T20:44:22.696Z" },
+    { url = "https://files.pythonhosted.org/packages/33/1c/c93b65d0b32e202596a2647922a75c7011cb982f899ddfcfd171f792c58f/onyx_devtools-0.6.3-py3-none-manylinux_2_17_aarch64.whl", hash = "sha256:384ef66030b55c0fd68b3898782b5b4b868ff3de119569dfc8544e2ce534b98a", size = 3540890, upload-time = "2026-03-05T20:39:28.886Z" },
+    { url = "https://files.pythonhosted.org/packages/f4/33/760eb656013f7f0cdff24570480d3dc4e52bbd8e6147ea1e8cf6fad7554f/onyx_devtools-0.6.3-py3-none-manylinux_2_17_x86_64.whl", hash = "sha256:82e218f3a49f64910c2c4c34d5dc12d1ea1520a27e0b0f6e4c0949ff9abaf0e1", size = 3945396, upload-time = "2026-03-05T20:39:34.323Z" },
+    { url = "https://files.pythonhosted.org/packages/1a/eb/f54b3675c464df8a51194ff75afc97c2417659e3a209dc46948b47c28860/onyx_devtools-0.6.3-py3-none-win_amd64.whl", hash = "sha256:8af614ae7229290ef2417cb85270184a1e826ed9a3a34658da93851edb36df57", size = 4045936, upload-time = "2026-03-05T20:39:28.375Z" },
+    { url = "https://files.pythonhosted.org/packages/04/b8/5bee38e748f3d4b8ec935766224db1bbc1214c91092e5822c080fccd9130/onyx_devtools-0.6.3-py3-none-win_arm64.whl", hash = "sha256:717589db4b42528d33ae96f8006ee6aad3555034dcfee724705b6576be6a6ec4", size = 3608268, upload-time = "2026-03-05T20:39:28.731Z" },
 ]
 
 [[package]]
@@ -5712,11 +5713,11 @@ wheels = [
 
 [[package]]
 name = "pypdf"
-version = "6.8.0"
+version = "6.7.5"
 source = { registry = "https://pypi.org/simple" }
-sdist = { url = "https://files.pythonhosted.org/packages/b4/a3/e705b0805212b663a4c27b861c8a603dba0f8b4bb281f96f8e746576a50d/pypdf-6.8.0.tar.gz", hash = "sha256:cb7eaeaa4133ce76f762184069a854e03f4d9a08568f0e0623f7ea810407833b", size = 5307831, upload-time = "2026-03-09T13:37:40.591Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/f6/52/37cc0aa9e9d1bf7729a737a0d83f8b3f851c8eb137373d9f71eafb0a3405/pypdf-6.7.5.tar.gz", hash = "sha256:40bb2e2e872078655f12b9b89e2f900888bb505e88a82150b64f9f34fa25651d", size = 5304278, upload-time = "2026-03-02T09:05:21.464Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/8c/ec/4ccf3bb86b1afe5d7176e1c8abcdbf22b53dd682ec2eda50e1caadcf6846/pypdf-6.8.0-py3-none-any.whl", hash = "sha256:2a025080a8dd73f48123c89c57174a5ff3806c71763ee4e49572dc90454943c7", size = 332177, upload-time = "2026-03-09T13:37:38.774Z" },
+    { url = "https://files.pythonhosted.org/packages/05/89/336673efd0a88956562658aba4f0bbef7cb92a6fbcbcaf94926dbc82b408/pypdf-6.7.5-py3-none-any.whl", hash = "sha256:07ba7f1d6e6d9aa2a17f5452e320a84718d4ce863367f7ede2fd72280349ab13", size = 331421, upload-time = "2026-03-02T09:05:19.722Z" },
 ]
 
 [[package]]
@@ -6337,16 +6338,16 @@ wheels = [
 
 [[package]]
 name = "release-tag"
-version = "0.5.2"
+version = "0.4.3"
 source = { registry = "https://pypi.org/simple" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/ab/92/01192a540b29cfadaa23850c8f6a2041d541b83a3fa1dc52a5f55212b3b6/release_tag-0.5.2-py3-none-any.whl", hash = "sha256:1e9ca7618bcfc63ad7a0728c84bbad52ef82d07586c4cc11365b44ea8f588069", size = 1264752, upload-time = "2026-03-11T00:27:18.674Z" },
-    { url = "https://files.pythonhosted.org/packages/4f/77/81fb42a23cd0de61caf84266f7aac1950b1c324883788b7c48e5344f61ae/release_tag-0.5.2-py3-none-macosx_10_12_x86_64.whl", hash = "sha256:8fbc61ff7bac2b96fab09566ec45c6508c201efc3f081f57702e1761bbc178d5", size = 1255075, upload-time = "2026-03-11T00:27:24.442Z" },
-    { url = "https://files.pythonhosted.org/packages/98/e6/769f8be94304529c1a531e995f2f3ac83f3c54738ce488b0abde75b20851/release_tag-0.5.2-py3-none-macosx_11_0_arm64.whl", hash = "sha256:fa3d7e495a0c516858a81878d03803539712677a3d6e015503de21cce19bea5e", size = 1163627, upload-time = "2026-03-11T00:27:26.412Z" },
-    { url = "https://files.pythonhosted.org/packages/45/68/7543e9daa0dfd41c487bf140d91fd5879327bb7c001a96aa5264667c30a1/release_tag-0.5.2-py3-none-manylinux_2_17_aarch64.whl", hash = "sha256:e8b60453218d6926da1fdcb99c2e17c851be0d7ab1975e97951f0bff5f32b565", size = 1140133, upload-time = "2026-03-11T00:27:20.633Z" },
-    { url = "https://files.pythonhosted.org/packages/6a/30/9087825696271012d889d136310dbdf0811976ae2b2f5a490f4e437903e1/release_tag-0.5.2-py3-none-manylinux_2_17_x86_64.whl", hash = "sha256:0e302ed60c2bf8b7ba5634842be28a27d83cec995869e112b0348b3f01a84ff5", size = 1264767, upload-time = "2026-03-11T00:27:28.355Z" },
-    { url = "https://files.pythonhosted.org/packages/79/a3/5b51b0cbdbf2299f545124beab182cfdfe01bf5b615efbc94aee3a64ea67/release_tag-0.5.2-py3-none-win_amd64.whl", hash = "sha256:e3c0629d373a16b9a3da965e89fca893640ce9878ec548865df3609b70989a89", size = 1340816, upload-time = "2026-03-11T00:27:22.622Z" },
-    { url = "https://files.pythonhosted.org/packages/dd/6f/832c2023a8bd8414c93452bd8b43bf61cedfa5b9575f70c06fb911e51a29/release_tag-0.5.2-py3-none-win_arm64.whl", hash = "sha256:5f26b008e0be0c7a122acd8fcb1bb5c822f38e77fed0c0bf6c550cc226c6bf14", size = 1203191, upload-time = "2026-03-11T00:27:29.789Z" },
+    { url = "https://files.pythonhosted.org/packages/39/18/c1d17d973f73f0aa7e2c45f852839ab909756e1bd9727d03babe400fcef0/release_tag-0.4.3-py3-none-any.whl", hash = "sha256:4206f4fa97df930c8176bfee4d3976a7385150ed14b317bd6bae7101ac8b66dd", size = 1181112, upload-time = "2025-12-03T00:18:19.445Z" },
+    { url = "https://files.pythonhosted.org/packages/33/c7/ecc443953840ac313856b2181f55eb8d34fa2c733cdd1edd0bcceee0938d/release_tag-0.4.3-py3-none-macosx_10_12_x86_64.whl", hash = "sha256:7a347a9ad3d2af16e5367e52b451fbc88a0b7b666850758e8f9a601554a8fb13", size = 1170517, upload-time = "2025-12-03T00:18:11.663Z" },
+    { url = "https://files.pythonhosted.org/packages/ce/81/2f6ffa0d87c792364ca9958433fe088c8acc3d096ac9734040049c6ad506/release_tag-0.4.3-py3-none-macosx_11_0_arm64.whl", hash = "sha256:2d1603aa37d8e4f5df63676bbfddc802fbc108a744ba28288ad25c997981c164", size = 1101663, upload-time = "2025-12-03T00:18:15.173Z" },
+    { url = "https://files.pythonhosted.org/packages/7c/ed/9e4ebe400fc52e38dda6e6a45d9da9decd4535ab15e170b8d9b229a66730/release_tag-0.4.3-py3-none-manylinux_2_17_aarch64.whl", hash = "sha256:6db7b81a198e3ba6a87496a554684912c13f9297ea8db8600a80f4f971709d37", size = 1079322, upload-time = "2025-12-03T00:18:16.094Z" },
+    { url = "https://files.pythonhosted.org/packages/2a/64/9e0ce6119e091ef9211fa82b9593f564eeec8bdd86eff6a97fe6e2fcb20f/release_tag-0.4.3-py3-none-manylinux_2_17_x86_64.whl", hash = "sha256:d79a9cf191dd2c29e1b3a35453fa364b08a7aadd15aeb2c556a7661c6cf4d5ad", size = 1181129, upload-time = "2025-12-03T00:18:15.82Z" },
+    { url = "https://files.pythonhosted.org/packages/b8/09/d96acf18f0773b6355080a568ba48931faa9dbe91ab1abefc6f8c4df04a8/release_tag-0.4.3-py3-none-win_amd64.whl", hash = "sha256:3958b880375f2241d0cc2b9882363bf54b1d4d7ca8ffc6eecc63ab92f23307f0", size = 1260773, upload-time = "2025-12-03T00:18:14.723Z" },
+    { url = "https://files.pythonhosted.org/packages/51/da/ecb6346df1ffb0752fe213e25062f802c10df2948717f0d5f9816c2df914/release_tag-0.4.3-py3-none-win_arm64.whl", hash = "sha256:7d5b08000e6e398d46f05a50139031046348fba6d47909f01e468bb7600c19df", size = 1142155, upload-time = "2025-12-03T00:18:20.647Z" },
 ]
 
 [[package]]
@@ -7232,19 +7233,21 @@ wheels = [
 
 [[package]]
 name = "tornado"
-version = "6.5.5"
+version = "6.5.2"
 source = { registry = "https://pypi.org/simple" }
-sdist = { url = "https://files.pythonhosted.org/packages/f8/f1/3173dfa4a18db4a9b03e5d55325559dab51ee653763bb8745a75af491286/tornado-6.5.5.tar.gz", hash = "sha256:192b8f3ea91bd7f1f50c06955416ed76c6b72f96779b962f07f911b91e8d30e9", size = 516006, upload-time = "2026-03-10T21:31:02.067Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/09/ce/1eb500eae19f4648281bb2186927bb062d2438c2e5093d1360391afd2f90/tornado-6.5.2.tar.gz", hash = "sha256:ab53c8f9a0fa351e2c0741284e06c7a45da86afb544133201c5cc8578eb076a0", size = 510821, upload-time = "2025-08-08T18:27:00.78Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/59/8c/77f5097695f4dd8255ecbd08b2a1ed8ba8b953d337804dd7080f199e12bf/tornado-6.5.5-cp39-abi3-macosx_10_9_universal2.whl", hash = "sha256:487dc9cc380e29f58c7ab88f9e27cdeef04b2140862e5076a66fb6bb68bb1bfa", size = 445983, upload-time = "2026-03-10T21:30:44.28Z" },
-    { url = "https://files.pythonhosted.org/packages/ab/5e/7625b76cd10f98f1516c36ce0346de62061156352353ef2da44e5c21523c/tornado-6.5.5-cp39-abi3-macosx_10_9_x86_64.whl", hash = "sha256:65a7f1d46d4bb41df1ac99f5fcb685fb25c7e61613742d5108b010975a9a6521", size = 444246, upload-time = "2026-03-10T21:30:46.571Z" },
-    { url = "https://files.pythonhosted.org/packages/b2/04/7b5705d5b3c0fab088f434f9c83edac1573830ca49ccf29fb83bf7178eec/tornado-6.5.5-cp39-abi3-manylinux1_x86_64.manylinux_2_28_x86_64.manylinux_2_5_x86_64.whl", hash = "sha256:e74c92e8e65086b338fd56333fb9a68b9f6f2fe7ad532645a290a464bcf46be5", size = 447229, upload-time = "2026-03-10T21:30:48.273Z" },
-    { url = "https://files.pythonhosted.org/packages/34/01/74e034a30ef59afb4097ef8659515e96a39d910b712a89af76f5e4e1f93c/tornado-6.5.5-cp39-abi3-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:435319e9e340276428bbdb4e7fa732c2d399386d1de5686cb331ec8eee754f07", size = 448192, upload-time = "2026-03-10T21:30:51.22Z" },
-    { url = "https://files.pythonhosted.org/packages/be/00/fe9e02c5a96429fce1a1d15a517f5d8444f9c412e0bb9eadfbe3b0fc55bf/tornado-6.5.5-cp39-abi3-musllinux_1_2_aarch64.whl", hash = "sha256:3f54aa540bdbfee7b9eb268ead60e7d199de5021facd276819c193c0fb28ea4e", size = 448039, upload-time = "2026-03-10T21:30:53.52Z" },
-    { url = "https://files.pythonhosted.org/packages/82/9e/656ee4cec0398b1d18d0f1eb6372c41c6b889722641d84948351ae19556d/tornado-6.5.5-cp39-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:36abed1754faeb80fbd6e64db2758091e1320f6bba74a4cf8c09cd18ccce8aca", size = 447445, upload-time = "2026-03-10T21:30:55.541Z" },
-    { url = "https://files.pythonhosted.org/packages/5a/76/4921c00511f88af86a33de770d64141170f1cfd9c00311aea689949e274e/tornado-6.5.5-cp39-abi3-win32.whl", hash = "sha256:dd3eafaaeec1c7f2f8fdcd5f964e8907ad788fe8a5a32c4426fbbdda621223b7", size = 448582, upload-time = "2026-03-10T21:30:57.142Z" },
-    { url = "https://files.pythonhosted.org/packages/2c/23/f6c6112a04d28eed765e374435fb1a9198f73e1ec4b4024184f21faeb1ad/tornado-6.5.5-cp39-abi3-win_amd64.whl", hash = "sha256:6443a794ba961a9f619b1ae926a2e900ac20c34483eea67be4ed8f1e58d3ef7b", size = 448990, upload-time = "2026-03-10T21:30:58.857Z" },
-    { url = "https://files.pythonhosted.org/packages/b7/c8/876602cbc96469911f0939f703453c1157b0c826ecb05bdd32e023397d4e/tornado-6.5.5-cp39-abi3-win_arm64.whl", hash = "sha256:2c9a876e094109333f888539ddb2de4361743e5d21eece20688e3e351e4990a6", size = 448016, upload-time = "2026-03-10T21:31:00.43Z" },
+    { url = "https://files.pythonhosted.org/packages/f6/48/6a7529df2c9cc12efd2e8f5dd219516184d703b34c06786809670df5b3bd/tornado-6.5.2-cp39-abi3-macosx_10_9_universal2.whl", hash = "sha256:2436822940d37cde62771cff8774f4f00b3c8024fe482e16ca8387b8a2724db6", size = 442563, upload-time = "2025-08-08T18:26:42.945Z" },
+    { url = "https://files.pythonhosted.org/packages/f2/b5/9b575a0ed3e50b00c40b08cbce82eb618229091d09f6d14bce80fc01cb0b/tornado-6.5.2-cp39-abi3-macosx_10_9_x86_64.whl", hash = "sha256:583a52c7aa94ee046854ba81d9ebb6c81ec0fd30386d96f7640c96dad45a03ef", size = 440729, upload-time = "2025-08-08T18:26:44.473Z" },
+    { url = "https://files.pythonhosted.org/packages/1b/4e/619174f52b120efcf23633c817fd3fed867c30bff785e2cd5a53a70e483c/tornado-6.5.2-cp39-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b0fe179f28d597deab2842b86ed4060deec7388f1fd9c1b4a41adf8af058907e", size = 444295, upload-time = "2025-08-08T18:26:46.021Z" },
+    { url = "https://files.pythonhosted.org/packages/95/fa/87b41709552bbd393c85dd18e4e3499dcd8983f66e7972926db8d96aa065/tornado-6.5.2-cp39-abi3-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:b186e85d1e3536d69583d2298423744740986018e393d0321df7340e71898882", size = 443644, upload-time = "2025-08-08T18:26:47.625Z" },
+    { url = "https://files.pythonhosted.org/packages/f9/41/fb15f06e33d7430ca89420283a8762a4e6b8025b800ea51796ab5e6d9559/tornado-6.5.2-cp39-abi3-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e792706668c87709709c18b353da1f7662317b563ff69f00bab83595940c7108", size = 443878, upload-time = "2025-08-08T18:26:50.599Z" },
+    { url = "https://files.pythonhosted.org/packages/11/92/fe6d57da897776ad2e01e279170ea8ae726755b045fe5ac73b75357a5a3f/tornado-6.5.2-cp39-abi3-musllinux_1_2_aarch64.whl", hash = "sha256:06ceb1300fd70cb20e43b1ad8aaee0266e69e7ced38fa910ad2e03285009ce7c", size = 444549, upload-time = "2025-08-08T18:26:51.864Z" },
+    { url = "https://files.pythonhosted.org/packages/9b/02/c8f4f6c9204526daf3d760f4aa555a7a33ad0e60843eac025ccfd6ff4a93/tornado-6.5.2-cp39-abi3-musllinux_1_2_i686.whl", hash = "sha256:74db443e0f5251be86cbf37929f84d8c20c27a355dd452a5cfa2aada0d001ec4", size = 443973, upload-time = "2025-08-08T18:26:53.625Z" },
+    { url = "https://files.pythonhosted.org/packages/ae/2d/f5f5707b655ce2317190183868cd0f6822a1121b4baeae509ceb9590d0bd/tornado-6.5.2-cp39-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:b5e735ab2889d7ed33b32a459cac490eda71a1ba6857b0118de476ab6c366c04", size = 443954, upload-time = "2025-08-08T18:26:55.072Z" },
+    { url = "https://files.pythonhosted.org/packages/e8/59/593bd0f40f7355806bf6573b47b8c22f8e1374c9b6fd03114bd6b7a3dcfd/tornado-6.5.2-cp39-abi3-win32.whl", hash = "sha256:c6f29e94d9b37a95013bb669616352ddb82e3bfe8326fccee50583caebc8a5f0", size = 445023, upload-time = "2025-08-08T18:26:56.677Z" },
+    { url = "https://files.pythonhosted.org/packages/c7/2a/f609b420c2f564a748a2d80ebfb2ee02a73ca80223af712fca591386cafb/tornado-6.5.2-cp39-abi3-win_amd64.whl", hash = "sha256:e56a5af51cc30dd2cae649429af65ca2f6571da29504a07995175df14c18f35f", size = 445427, upload-time = "2025-08-08T18:26:57.91Z" },
+    { url = "https://files.pythonhosted.org/packages/5e/4f/e1f65e8f8c76d73658b33d33b81eed4322fb5085350e4328d5c956f0c8f9/tornado-6.5.2-cp39-abi3-win_arm64.whl", hash = "sha256:d6c33dc3672e3a1f3618eb63b7ef4683a7688e7b9e6e8f0d9aa5726360a004af", size = 444456, upload-time = "2025-08-08T18:26:59.207Z" },
 ]
 
 [[package]]

web/jest.config.js

@@ -143,9 +143,7 @@ module.exports = {
         "**/src/app/**/utils/*.test.ts",
         "**/src/app/**/hooks/*.test.ts", // Pure packet processor tests
         "**/src/refresh-components/**/*.test.ts",
-        "**/src/refresh-pages/**/*.test.ts",
         "**/src/sections/**/*.test.ts",
-        "**/src/components/**/*.test.ts",
         // Add more patterns here as you add more unit tests
       ],
     },

web/lib/opal/src/components/buttons/OpenButton/OpenButton.stories.tsx

@@ -1,6 +1,5 @@
 import type { Meta, StoryObj } from "@storybook/react";
 import { OpenButton } from "@opal/components";
-import { Disabled as DisabledProvider } from "@opal/core";
 import { SvgSettings } from "@opal/icons";
 import * as TooltipPrimitive from "@radix-ui/react-tooltip";
 
@@ -33,9 +32,16 @@ export const WithIcon: Story = {
   },
 };
 
+export const Selected: Story = {
+  args: {
+    selected: true,
+    children: "Selected",
+  },
+};
+
 export const Open: Story = {
   args: {
-    interaction: "hover",
+    transient: true,
     children: "Open state",
   },
 };
@@ -47,27 +53,18 @@ export const Disabled: Story = {
   },
 };
 
-export const Foldable: Story = {
+export const LightProminence: Story = {
   args: {
-    foldable: true,
-    icon: SvgSettings,
-    children: "Settings",
+    prominence: "light",
+    children: "Light prominence",
   },
 };
 
-export const FoldableDisabled: Story = {
+export const HeavyProminence: Story = {
   args: {
-    foldable: true,
-    icon: SvgSettings,
-    children: "Settings",
+    prominence: "heavy",
+    children: "Heavy prominence",
   },
-  decorators: [
-    (Story) => (
-      <DisabledProvider disabled>
-        <Story />
-      </DisabledProvider>
-    ),
-  ],
 };
 
 export const Sizes: Story = {
@@ -81,12 +78,3 @@ export const Sizes: Story = {
     </div>
   ),
 };
-
-export const WithTooltip: Story = {
-  args: {
-    icon: SvgSettings,
-    children: "Settings",
-    tooltip: "Open settings",
-    tooltipSide: "bottom",
-  },
-};

web/lib/opal/src/components/buttons/open-button/README.md

@@ -17,9 +17,7 @@ OpenButton is a **tighter, specialized use-case** of SelectButton:
 - It hardcodes `variant="select-heavy"` (SelectButton exposes `variant`)
 - It adds a built-in chevron with CSS-driven rotation (SelectButton has no chevron)
 - It auto-detects Radix `data-state="open"` to derive `interaction` (SelectButton has no Radix awareness)
-- It does not support `rightIcon` (SelectButton does)
-
-Both components support `foldable` using the same pattern: `interactive-foldable-host` class + `Interactive.Foldable` wrapper around the label and trailing icon. When foldable, the left icon stays visible while the rest collapses. If you change the foldable implementation in one, update the other to match.
+- It does not support `foldable` or `rightIcon` (SelectButton does)
 
 If you need a general-purpose stateful toggle, use `SelectButton`. If you need a popover/dropdown trigger with a chevron, use `OpenButton`.
 
@@ -28,12 +26,10 @@ If you need a general-purpose stateful toggle, use `SelectButton`. If you need a
 ```
 Interactive.Stateful           <- variant="select-heavy", interaction, state, disabled, onClick
   └─ Interactive.Container     <- height, rounding, padding (from `size`)
-       └─ div.opal-button.interactive-foreground [.interactive-foldable-host]
+       └─ div.opal-button.interactive-foreground
             ├─ div > Icon?                 (interactive-foreground-icon)
-            ├─ [Foldable]?                 (wraps label + chevron when foldable)
-            │    ├─ <span>?                .opal-button-label
-            │    └─ div > ChevronIcon      .opal-open-button-chevron
-            └─ <span>? / ChevronIcon       (non-foldable)
+            ├─ <span>?                     .opal-button-label
+            └─ div > ChevronIcon           .opal-open-button-chevron (interactive-foreground-icon)
 ```
 
 - **`interaction` controls both the chevron and the hover visual state.** When `interaction` is `"hover"` (explicitly or via Radix `data-state="open"`), the chevron rotates 180° and the hover background activates.
@@ -48,7 +44,6 @@ Interactive.Stateful           <- variant="select-heavy", interaction, state, di
 | `interaction` | `"rest" \| "hover" \| "active"` | auto | JS-controlled interaction override. Falls back to Radix `data-state="open"` when omitted. |
 | `icon` | `IconFunctionComponent` | — | Left icon component |
 | `children` | `string` | — | Content between icon and chevron |
-| `foldable` | `boolean` | `false` | When `true`, requires both `icon` and `children`; the left icon stays visible while the label + chevron collapse when not hovered. If `tooltip` is omitted on a disabled foldable button, the label text is used as the tooltip. |
 | `size` | `SizeVariant` | `"lg"` | Size preset controlling height, rounding, and padding |
 | `width` | `WidthVariant` | — | Width preset |
 | `tooltip` | `string` | — | Tooltip text shown on hover |

web/lib/opal/src/components/buttons/open-button/components.tsx

@@ -2,7 +2,6 @@ import "@opal/components/buttons/open-button/styles.css";
 import "@opal/components/tooltip.css";
 import {
   Interactive,
-  useDisabled,
   type InteractiveStatefulProps,
   type InteractiveStatefulInteraction,
 } from "@opal/core";
@@ -31,46 +30,27 @@ function ChevronIcon({ className, ...props }: IconProps) {
 // Types
 // ---------------------------------------------------------------------------
 
-/**
- * Content props — a discriminated union on `foldable` that enforces:
- *
- * - `foldable: true`  → `icon` and `children` are required (icon stays visible,
- *                        label + chevron fold away)
- * - `foldable?: false` → at least one of `icon` or `children` must be provided
- */
-type OpenButtonContentProps =
-  | {
-      foldable: true;
-      icon: IconFunctionComponent;
-      children: string;
-    }
-  | {
-      foldable?: false;
-      icon?: IconFunctionComponent;
-      children: string;
-    }
-  | {
-      foldable?: false;
-      icon: IconFunctionComponent;
-      children?: string;
-    };
+type OpenButtonProps = Omit<InteractiveStatefulProps, "variant"> & {
+  /** Left icon. */
+  icon?: IconFunctionComponent;
 
-type OpenButtonProps = Omit<InteractiveStatefulProps, "variant"> &
-  OpenButtonContentProps & {
-    /**
-     * Size preset — controls gap, text size, and Container height/rounding.
-     */
-    size?: SizeVariant;
+  /** Button label text. */
+  children?: string;
 
-    /** Width preset. */
-    width?: WidthVariant;
+  /**
+   * Size preset — controls gap, text size, and Container height/rounding.
+   */
+  size?: SizeVariant;
 
-    /** Tooltip text shown on hover. */
-    tooltip?: string;
+  /** Width preset. */
+  width?: WidthVariant;
 
-    /** Which side the tooltip appears on. */
-    tooltipSide?: TooltipSide;
-  };
+  /** Tooltip text shown on hover. */
+  tooltip?: string;
+
+  /** Which side the tooltip appears on. */
+  tooltipSide?: TooltipSide;
+};
 
 // ---------------------------------------------------------------------------
 // OpenButton
@@ -80,15 +60,12 @@ function OpenButton({
   icon: Icon,
   children,
   size = "lg",
-  foldable,
   width,
   tooltip,
   tooltipSide = "top",
   interaction,
   ...statefulProps
 }: OpenButtonProps) {
-  const { isDisabled } = useDisabled();
-
   // Derive open state: explicit prop → Radix data-state (injected via Slot chain)
   const dataState = (statefulProps as Record<string, unknown>)["data-state"] as
     | string
@@ -98,17 +75,6 @@ function OpenButton({
 
   const isLarge = size === "lg";
 
-  const labelEl = children ? (
-    <span
-      className={cn(
-        "opal-button-label whitespace-nowrap",
-        isLarge ? "font-main-ui-body" : "font-secondary-body"
-      )}
-    >
-      {children}
-    </span>
-  ) : null;
-
   const button = (
     <Interactive.Stateful
       variant="select-heavy"
@@ -123,34 +89,25 @@ function OpenButton({
           isLarge ? "default" : size === "2xs" ? "mini" : "compact"
         }
       >
-        <div
-          className={cn(
-            "opal-button interactive-foreground flex flex-row items-center gap-1",
-            foldable && "interactive-foldable-host"
-          )}
-        >
-          {iconWrapper(Icon, size, !foldable && !!children)}
-
-          {foldable ? (
-            <Interactive.Foldable>
-              {labelEl}
-              {iconWrapper(ChevronIcon, size, !!children)}
-            </Interactive.Foldable>
-          ) : (
-            <>
-              {labelEl}
-              {iconWrapper(ChevronIcon, size, !!children)}
-            </>
+        <div className="opal-button interactive-foreground flex flex-row items-center gap-1">
+          {iconWrapper(Icon, size, false)}
+          {children && (
+            <span
+              className={cn(
+                "opal-button-label whitespace-nowrap",
+                isLarge ? "font-main-ui-body" : "font-secondary-body"
+              )}
+            >
+              {children}
+            </span>
           )}
+          {iconWrapper(ChevronIcon, size, false)}
         </div>
       </Interactive.Container>
     </Interactive.Stateful>
   );
 
-  const resolvedTooltip =
-    tooltip ?? (foldable && isDisabled && children ? children : undefined);
-
-  if (!resolvedTooltip) return button;
+  if (!tooltip) return button;
 
   return (
     <TooltipPrimitive.Root>
@@ -161,7 +118,7 @@ function OpenButton({
           side={tooltipSide}
           sideOffset={4}
         >
-          {resolvedTooltip}
+          {tooltip}
         </TooltipPrimitive.Content>
       </TooltipPrimitive.Portal>
     </TooltipPrimitive.Root>

web/lib/opal/src/components/buttons/select-button/README.md

@@ -17,9 +17,7 @@ Interactive.Stateful → Interactive.Container → content row (icon + label + t
 - OpenButton hardcodes `variant="select-heavy"` (SelectButton exposes `variant`)
 - OpenButton adds a built-in chevron with CSS-driven rotation (SelectButton has no chevron)
 - OpenButton auto-detects Radix `data-state="open"` to derive `interaction` (SelectButton has no Radix awareness)
-- OpenButton does not support `rightIcon` (SelectButton does)
-
-Both components support `foldable` using the same pattern: `interactive-foldable-host` class + `Interactive.Foldable` wrapper around the label and trailing icon. When foldable, the left icon stays visible while the rest collapses. If you change the foldable implementation in one, update the other to match.
+- OpenButton does not support `foldable` or `rightIcon` (SelectButton does)
 
 Use SelectButton for general-purpose stateful toggles. Use `OpenButton` for popover/dropdown triggers with a chevron.
 

web/lib/opal/src/components/cards/card/Card.stories.tsx

@@ -1,87 +0,0 @@
-import type { Meta, StoryObj } from "@storybook/react";
-import { Card } from "@opal/components";
-
-const BACKGROUND_VARIANTS = ["none", "light", "heavy"] as const;
-const BORDER_VARIANTS = ["none", "dashed", "solid"] as const;
-const SIZE_VARIANTS = ["lg", "md", "sm", "xs", "2xs", "fit"] as const;
-
-const meta: Meta<typeof Card> = {
-  title: "opal/components/Card",
-  component: Card,
-  tags: ["autodocs"],
-};
-
-export default meta;
-type Story = StoryObj<typeof Card>;
-
-export const Default: Story = {
-  render: () => (
-    <Card>
-      <p>Default card with light background, no border, lg size.</p>
-    </Card>
-  ),
-};
-
-export const BackgroundVariants: Story = {
-  render: () => (
-    <div className="flex flex-col gap-4 w-96">
-      {BACKGROUND_VARIANTS.map((bg) => (
-        <Card key={bg} backgroundVariant={bg} borderVariant="solid">
-          <p>backgroundVariant: {bg}</p>
-        </Card>
-      ))}
-    </div>
-  ),
-};
-
-export const BorderVariants: Story = {
-  render: () => (
-    <div className="flex flex-col gap-4 w-96">
-      {BORDER_VARIANTS.map((border) => (
-        <Card key={border} borderVariant={border}>
-          <p>borderVariant: {border}</p>
-        </Card>
-      ))}
-    </div>
-  ),
-};
-
-export const SizeVariants: Story = {
-  render: () => (
-    <div className="flex flex-col gap-4 w-96">
-      {SIZE_VARIANTS.map((size) => (
-        <Card key={size} sizeVariant={size} borderVariant="solid">
-          <p>sizeVariant: {size}</p>
-        </Card>
-      ))}
-    </div>
-  ),
-};
-
-export const AllCombinations: Story = {
-  render: () => (
-    <div className="flex flex-col gap-8">
-      {SIZE_VARIANTS.map((size) => (
-        <div key={size}>
-          <p className="font-bold pb-2">sizeVariant: {size}</p>
-          <div className="grid grid-cols-3 gap-4">
-            {BACKGROUND_VARIANTS.map((bg) =>
-              BORDER_VARIANTS.map((border) => (
-                <Card
-                  key={`${size}-${bg}-${border}`}
-                  sizeVariant={size}
-                  backgroundVariant={bg}
-                  borderVariant={border}
-                >
-                  <p className="text-xs">
-                    bg: {bg}, border: {border}
-                  </p>
-                </Card>
-              ))
-            )}
-          </div>
-        </div>
-      ))}
-    </div>
-  ),
-};

web/lib/opal/src/components/cards/card/README.md

@@ -1,67 +0,0 @@
-# Card
-
-**Import:** `import { Card, type CardProps } from "@opal/components";`
-
-A plain container component with configurable background, border, padding, and rounding. Uses a simple `<div>` internally with `overflow-clip`.
-
-## Architecture
-
-The `sizeVariant` controls both padding and border-radius, mirroring the same mapping used by `Button` and `Interactive.Container`:
-
-| Size      | Padding | Rounding       |
-|-----------|---------|----------------|
-| `lg`      | `p-2`   | `rounded-12`   |
-| `md`      | `p-1`   | `rounded-08`   |
-| `sm`      | `p-1`   | `rounded-08`   |
-| `xs`      | `p-0.5` | `rounded-04`   |
-| `2xs`     | `p-0.5` | `rounded-04`   |
-| `fit`     | `p-0`   | `rounded-12`   |
-
-## Props
-
-| Prop | Type | Default | Description |
-|------|------|---------|-------------|
-| `sizeVariant` | `SizeVariant` | `"lg"` | Controls padding and border-radius |
-| `backgroundVariant` | `"none" \| "light" \| "heavy"` | `"light"` | Background fill intensity |
-| `borderVariant` | `"none" \| "dashed" \| "solid"` | `"none"` | Border style |
-| `ref` | `React.Ref<HTMLDivElement>` | — | Ref forwarded to the root div |
-| `children` | `React.ReactNode` | — | Card content |
-
-## Background Variants
-
-- **`none`** — Transparent background. Use for seamless inline content.
-- **`light`** — Subtle tinted background (`bg-background-tint-00`). The default, suitable for most cards.
-- **`heavy`** — Stronger tinted background (`bg-background-tint-01`). Use for emphasis or nested cards that need visual separation.
-
-## Border Variants
-
-- **`none`** — No border. Use when cards are visually grouped or in tight layouts.
-- **`dashed`** — Dashed border. Use for placeholder or empty states.
-- **`solid`** — Solid border. Use for prominent, standalone cards.
-
-## Usage
-
-```tsx
-import { Card } from "@opal/components";
-
-// Default card (light background, no border, lg padding + rounding)
-<Card>
-  <h2>Card Title</h2>
-  <p>Card content</p>
-</Card>
-
-// Compact card with solid border
-<Card borderVariant="solid" sizeVariant="sm">
-  <p>Compact card</p>
-</Card>
-
-// Empty state card
-<Card backgroundVariant="none" borderVariant="dashed">
-  <p>No items yet</p>
-</Card>
-
-// Heavy background, tight padding
-<Card backgroundVariant="heavy" sizeVariant="xs">
-  <p>Highlighted content</p>
-</Card>
-```

web/lib/opal/src/components/cards/card/components.tsx

@@ -1,101 +0,0 @@
-import "@opal/components/cards/card/styles.css";
-import type { SizeVariant } from "@opal/shared";
-import { sizeVariants } from "@opal/shared";
-import { cn } from "@opal/utils";
-
-// ---------------------------------------------------------------------------
-// Types
-// ---------------------------------------------------------------------------
-
-type BackgroundVariant = "none" | "light" | "heavy";
-type BorderVariant = "none" | "dashed" | "solid";
-
-type CardProps = {
-  /**
-   * Size preset — controls padding and border-radius.
-   *
-   * Padding comes from the shared size scale. Rounding follows the same
-   * mapping as `Button` / `Interactive.Container`:
-   *
-   * | Size   | Rounding   |
-   * |--------|------------|
-   * | `lg`   | `default`  |
-   * | `md`–`sm` | `compact` |
-   * | `xs`–`2xs` | `mini`  |
-   * | `fit`  | `default`  |
-   *
-   * @default "lg"
-   */
-  sizeVariant?: SizeVariant;
-
-  /**
-   * Background fill intensity.
-   * - `"none"`: transparent background.
-   * - `"light"`: subtle tinted background (`bg-background-tint-00`).
-   * - `"heavy"`: stronger tinted background (`bg-background-tint-01`).
-   *
-   * @default "light"
-   */
-  backgroundVariant?: BackgroundVariant;
-
-  /**
-   * Border style.
-   * - `"none"`: no border.
-   * - `"dashed"`: dashed border.
-   * - `"solid"`: solid border.
-   *
-   * @default "none"
-   */
-  borderVariant?: BorderVariant;
-
-  /** Ref forwarded to the root `<div>`. */
-  ref?: React.Ref<HTMLDivElement>;
-
-  children?: React.ReactNode;
-};
-
-// ---------------------------------------------------------------------------
-// Rounding
-// ---------------------------------------------------------------------------
-
-/** Maps a size variant to a rounding class, mirroring the Button pattern. */
-const roundingForSize: Record<SizeVariant, string> = {
-  lg: "rounded-12",
-  md: "rounded-08",
-  sm: "rounded-08",
-  xs: "rounded-04",
-  "2xs": "rounded-04",
-  fit: "rounded-12",
-};
-
-// ---------------------------------------------------------------------------
-// Card
-// ---------------------------------------------------------------------------
-
-function Card({
-  sizeVariant = "lg",
-  backgroundVariant = "light",
-  borderVariant = "none",
-  ref,
-  children,
-}: CardProps) {
-  const { padding } = sizeVariants[sizeVariant];
-  const rounding = roundingForSize[sizeVariant];
-
-  return (
-    <div
-      ref={ref}
-      className={cn("opal-card", padding, rounding)}
-      data-background={backgroundVariant}
-      data-border={borderVariant}
-    >
-      {children}
-    </div>
-  );
-}
-
-// ---------------------------------------------------------------------------
-// Exports
-// ---------------------------------------------------------------------------
-
-export { Card, type CardProps, type BackgroundVariant, type BorderVariant };

web/lib/opal/src/components/cards/card/styles.css

@@ -1,29 +0,0 @@
-.opal-card {
-  @apply w-full overflow-clip;
-}
-
-/* Background variants */
-.opal-card[data-background="none"] {
-  @apply bg-transparent;
-}
-
-.opal-card[data-background="light"] {
-  @apply bg-background-tint-00;
-}
-
-.opal-card[data-background="heavy"] {
-  @apply bg-background-tint-01;
-}
-
-/* Border variants */
-.opal-card[data-border="none"] {
-  border: none;
-}
-
-.opal-card[data-border="dashed"] {
-  @apply border border-dashed;
-}
-
-.opal-card[data-border="solid"] {
-  @apply border;
-}

web/lib/opal/src/components/cards/empty-message-card/EmptyMessageCard.stories.tsx

@@ -1,51 +0,0 @@
-import type { Meta, StoryObj } from "@storybook/react";
-import { EmptyMessageCard } from "@opal/components";
-import { SvgSparkle, SvgUsers } from "@opal/icons";
-
-const SIZE_VARIANTS = ["lg", "md", "sm", "xs", "2xs", "fit"] as const;
-
-const meta: Meta<typeof EmptyMessageCard> = {
-  title: "opal/components/EmptyMessageCard",
-  component: EmptyMessageCard,
-  tags: ["autodocs"],
-};
-
-export default meta;
-type Story = StoryObj<typeof EmptyMessageCard>;
-
-export const Default: Story = {
-  args: {
-    title: "No items available.",
-  },
-};
-
-export const WithCustomIcon: Story = {
-  args: {
-    icon: SvgSparkle,
-    title: "No agents selected.",
-  },
-};
-
-export const SizeVariants: Story = {
-  render: () => (
-    <div className="flex flex-col gap-4 w-96">
-      {SIZE_VARIANTS.map((size) => (
-        <EmptyMessageCard
-          key={size}
-          sizeVariant={size}
-          title={`sizeVariant: ${size}`}
-        />
-      ))}
-    </div>
-  ),
-};
-
-export const Multiple: Story = {
-  render: () => (
-    <div className="flex flex-col gap-4 w-96">
-      <EmptyMessageCard title="No models available." />
-      <EmptyMessageCard icon={SvgSparkle} title="No agents selected." />
-      <EmptyMessageCard icon={SvgUsers} title="No groups added." />
-    </div>
-  ),
-};

web/lib/opal/src/components/cards/empty-message-card/README.md

@@ -1,30 +0,0 @@
-# EmptyMessageCard
-
-**Import:** `import { EmptyMessageCard, type EmptyMessageCardProps } from "@opal/components";`
-
-A pre-configured Card for empty states. Renders a transparent card with a dashed border containing a muted icon and message text using the `Content` layout.
-
-## Props
-
-| Prop          | Type                       | Default    | Description                                      |
-| ------------- | -------------------------- | ---------- | ------------------------------------------------ |
-| `icon`        | `IconFunctionComponent`    | `SvgEmpty` | Icon displayed alongside the title               |
-| `title`       | `string`                   | —          | Primary message text (required)                  |
-| `sizeVariant` | `SizeVariant`              | `"lg"`     | Size preset controlling padding and rounding     |
-| `ref`         | `React.Ref<HTMLDivElement>` | —          | Ref forwarded to the root div                    |
-
-## Usage
-
-```tsx
-import { EmptyMessageCard } from "@opal/components";
-import { SvgSparkle, SvgFileText } from "@opal/icons";
-
-// Default empty state
-<EmptyMessageCard title="No items yet." />
-
-// With custom icon
-<EmptyMessageCard icon={SvgSparkle} title="No agents selected." />
-
-// With custom size
-<EmptyMessageCard sizeVariant="sm" icon={SvgFileText} title="No documents available." />
-```

web/lib/opal/src/components/cards/empty-message-card/components.tsx

@@ -1,57 +0,0 @@
-import { Card } from "@opal/components/cards/card/components";
-import { Content } from "@opal/layouts";
-import { SvgEmpty } from "@opal/icons";
-import type { SizeVariant } from "@opal/shared";
-import type { IconFunctionComponent } from "@opal/types";
-
-// ---------------------------------------------------------------------------
-// Types
-// ---------------------------------------------------------------------------
-
-type EmptyMessageCardProps = {
-  /** Icon displayed alongside the title. */
-  icon?: IconFunctionComponent;
-
-  /** Primary message text. */
-  title: string;
-
-  /** Size preset controlling padding and rounding of the card. */
-  sizeVariant?: SizeVariant;
-
-  /** Ref forwarded to the root Card div. */
-  ref?: React.Ref<HTMLDivElement>;
-};
-
-// ---------------------------------------------------------------------------
-// EmptyMessageCard
-// ---------------------------------------------------------------------------
-
-function EmptyMessageCard({
-  icon = SvgEmpty,
-  title,
-  sizeVariant = "lg",
-  ref,
-}: EmptyMessageCardProps) {
-  return (
-    <Card
-      ref={ref}
-      backgroundVariant="none"
-      borderVariant="dashed"
-      sizeVariant={sizeVariant}
-    >
-      <Content
-        icon={icon}
-        title={title}
-        sizePreset="secondary"
-        variant="body"
-        prominence="muted"
-      />
-    </Card>
-  );
-}
-
-// ---------------------------------------------------------------------------
-// Exports
-// ---------------------------------------------------------------------------
-
-export { EmptyMessageCard, type EmptyMessageCardProps };

web/lib/opal/src/components/index.ts

@@ -31,17 +31,3 @@ export {
   type TagProps,
   type TagColor,
 } from "@opal/components/tag/components";
-
-/* Card */
-export {
-  Card,
-  type CardProps,
-  type BackgroundVariant,
-  type BorderVariant,
-} from "@opal/components/cards/card/components";
-
-/* EmptyMessageCard */
-export {
-  EmptyMessageCard,
-  type EmptyMessageCardProps,
-} from "@opal/components/cards/empty-message-card/components";

web/lib/opal/src/core/interactive/Interactive.stories.tsx

@@ -1,5 +1,5 @@
 import type { Meta, StoryObj } from "@storybook/react";
-import { Interactive, Disabled } from "@opal/core";
+import { Interactive } from "@opal/core";
 
 // ---------------------------------------------------------------------------
 // Variant / Prominence mappings for the matrix story
@@ -9,6 +9,8 @@ const VARIANT_PROMINENCE_MAP: Record<string, string[]> = {
   default: ["primary", "secondary", "tertiary", "internal"],
   action: ["primary", "secondary", "tertiary", "internal"],
   danger: ["primary", "secondary", "tertiary", "internal"],
+  select: ["light", "heavy"],
+  sidebar: ["light"],
   none: [],
 };
 
@@ -33,39 +35,39 @@ export default meta;
 // Stories
 // ---------------------------------------------------------------------------
 
-/** Basic Interactive.Stateless + Container with text content. */
+/** Basic Interactive.Base + Container with text content. */
 export const Default: StoryObj = {
   render: () => (
     <div style={{ display: "flex", gap: "0.75rem", alignItems: "center" }}>
-      <Interactive.Stateless
+      <Interactive.Base
         variant="default"
         prominence="secondary"
         onClick={() => {}}
       >
         <Interactive.Container border>
-          <span className="interactive-foreground">Secondary</span>
+          <span>Secondary</span>
         </Interactive.Container>
-      </Interactive.Stateless>
+      </Interactive.Base>
 
-      <Interactive.Stateless
+      <Interactive.Base
         variant="default"
         prominence="primary"
         onClick={() => {}}
       >
         <Interactive.Container border>
-          <span className="interactive-foreground">Primary</span>
+          <span>Primary</span>
         </Interactive.Container>
-      </Interactive.Stateless>
+      </Interactive.Base>
 
-      <Interactive.Stateless
+      <Interactive.Base
         variant="default"
         prominence="tertiary"
         onClick={() => {}}
       >
         <Interactive.Container border>
-          <span className="interactive-foreground">Tertiary</span>
+          <span>Tertiary</span>
         </Interactive.Container>
-      </Interactive.Stateless>
+      </Interactive.Base>
     </div>
   ),
 };
@@ -89,13 +91,11 @@ export const VariantMatrix: StoryObj = {
           </div>
 
           {prominences.length === 0 ? (
-            <Interactive.Stateless variant="none" onClick={() => {}}>
+            <Interactive.Base variant="none" onClick={() => {}}>
               <Interactive.Container border>
-                <span style={{ color: "var(--text-01)" }}>
-                  none (no prominence)
-                </span>
+                <span>none (no prominence)</span>
               </Interactive.Container>
-            </Interactive.Stateless>
+            </Interactive.Base>
           ) : (
             <div style={{ display: "flex", gap: "0.5rem", flexWrap: "wrap" }}>
               {prominences.map((prominence) => (
@@ -108,18 +108,16 @@ export const VariantMatrix: StoryObj = {
                     gap: "0.25rem",
                   }}
                 >
-                  <Interactive.Stateless
+                  <Interactive.Base
                     // Cast required because the discriminated union can't be
                     // resolved from dynamic strings at the type level.
                     {...({ variant, prominence } as any)}
                     onClick={() => {}}
                   >
                     <Interactive.Container border>
-                      <span className="interactive-foreground">
-                        {prominence}
-                      </span>
+                      <span>{prominence}</span>
                     </Interactive.Container>
-                  </Interactive.Stateless>
+                  </Interactive.Base>
                   <span
                     style={{
                       fontSize: "0.625rem",
@@ -143,16 +141,16 @@ export const Sizes: StoryObj = {
   render: () => (
     <div style={{ display: "flex", alignItems: "center", gap: "0.75rem" }}>
       {SIZE_VARIANTS.map((size) => (
-        <Interactive.Stateless
+        <Interactive.Base
           key={size}
           variant="default"
           prominence="secondary"
           onClick={() => {}}
         >
           <Interactive.Container border heightVariant={size}>
-            <span className="interactive-foreground">{size}</span>
+            <span>{size}</span>
           </Interactive.Container>
-        </Interactive.Stateless>
+        </Interactive.Base>
       ))}
     </div>
   ),
@@ -162,15 +160,15 @@ export const Sizes: StoryObj = {
 export const WidthFull: StoryObj = {
   render: () => (
     <div style={{ width: 400 }}>
-      <Interactive.Stateless
+      <Interactive.Base
         variant="default"
         prominence="secondary"
         onClick={() => {}}
       >
         <Interactive.Container border widthVariant="full">
-          <span className="interactive-foreground">Full width container</span>
+          <span>Full width container</span>
         </Interactive.Container>
-      </Interactive.Stateless>
+      </Interactive.Base>
     </div>
   ),
 };
@@ -180,86 +178,73 @@ export const Rounding: StoryObj = {
   render: () => (
     <div style={{ display: "flex", gap: "0.75rem" }}>
       {ROUNDING_VARIANTS.map((rounding) => (
-        <Interactive.Stateless
+        <Interactive.Base
           key={rounding}
           variant="default"
           prominence="secondary"
           onClick={() => {}}
         >
           <Interactive.Container border roundingVariant={rounding}>
-            <span className="interactive-foreground">{rounding}</span>
+            <span>{rounding}</span>
           </Interactive.Container>
-        </Interactive.Stateless>
+        </Interactive.Base>
       ))}
     </div>
   ),
 };
 
 /** Disabled state prevents clicks and shows disabled styling. */
-export const DisabledStory: StoryObj = {
-  name: "Disabled",
+export const Disabled: StoryObj = {
   render: () => (
     <div style={{ display: "flex", gap: "0.75rem" }}>
-      <Disabled disabled>
-        <Interactive.Stateless
-          variant="default"
-          prominence="secondary"
-          onClick={() => {}}
-        >
-          <Interactive.Container border>
-            <span className="interactive-foreground">Disabled</span>
-          </Interactive.Container>
-        </Interactive.Stateless>
-      </Disabled>
+      <Interactive.Base
+        variant="default"
+        prominence="secondary"
+        onClick={() => {}}
+        disabled
+      >
+        <Interactive.Container border>
+          <span>Disabled</span>
+        </Interactive.Container>
+      </Interactive.Base>
 
-      <Interactive.Stateless
+      <Interactive.Base
         variant="default"
         prominence="secondary"
         onClick={() => {}}
       >
         <Interactive.Container border>
-          <span className="interactive-foreground">Enabled</span>
+          <span>Enabled</span>
         </Interactive.Container>
-      </Interactive.Stateless>
+      </Interactive.Base>
     </div>
   ),
 };
 
-/** Interaction override forces the hover/active visual state. */
-export const Interaction: StoryObj = {
+/** Transient prop forces the hover/active visual state. */
+export const Transient: StoryObj = {
   render: () => (
     <div style={{ display: "flex", gap: "0.75rem" }}>
-      <Interactive.Stateless
+      <Interactive.Base
         variant="default"
         prominence="secondary"
-        interaction="hover"
         onClick={() => {}}
+        transient
       >
         <Interactive.Container border>
-          <span className="interactive-foreground">Forced hover</span>
+          <span>Forced hover</span>
         </Interactive.Container>
-      </Interactive.Stateless>
+      </Interactive.Base>
 
-      <Interactive.Stateless
-        variant="default"
-        prominence="secondary"
-        interaction="active"
-        onClick={() => {}}
-      >
-        <Interactive.Container border>
-          <span className="interactive-foreground">Forced active</span>
-        </Interactive.Container>
-      </Interactive.Stateless>
-
-      <Interactive.Stateless
+      <Interactive.Base
         variant="default"
         prominence="secondary"
         onClick={() => {}}
       >
         <Interactive.Container border>
-          <span className="interactive-foreground">Normal (rest)</span>
+          <span>Normal</span>
         </Interactive.Container>
-      </Interactive.Stateless>
+      </Interactive.Base>
     </div>
   ),
 };
@@ -268,25 +253,25 @@ export const Interaction: StoryObj = {
 export const WithBorder: StoryObj = {
   render: () => (
     <div style={{ display: "flex", gap: "0.75rem" }}>
-      <Interactive.Stateless
+      <Interactive.Base
         variant="default"
         prominence="secondary"
         onClick={() => {}}
       >
         <Interactive.Container border>
-          <span className="interactive-foreground">With border</span>
+          <span>With border</span>
         </Interactive.Container>
-      </Interactive.Stateless>
+      </Interactive.Base>
 
-      <Interactive.Stateless
+      <Interactive.Base
         variant="default"
         prominence="secondary"
         onClick={() => {}}
       >
         <Interactive.Container>
-          <span className="interactive-foreground">Without border</span>
+          <span>Without border</span>
         </Interactive.Container>
-      </Interactive.Stateless>
+      </Interactive.Base>
     </div>
   ),
 };
@@ -294,57 +279,51 @@ export const WithBorder: StoryObj = {
 /** Using href to render as a link. */
 export const AsLink: StoryObj = {
   render: () => (
-    <Interactive.Stateless variant="action" href="/settings">
+    <Interactive.Base variant="action" href="/settings">
       <Interactive.Container border>
-        <span className="interactive-foreground">Go to Settings</span>
+        <span>Go to Settings</span>
       </Interactive.Container>
-    </Interactive.Stateless>
+    </Interactive.Base>
   ),
 };
 
-/** Stateful select variant with selected and unselected states. */
+/** Select variant with selected and unselected states. */
 export const SelectVariant: StoryObj = {
   render: () => (
     <div style={{ display: "flex", gap: "0.75rem" }}>
-      <Interactive.Stateful
-        variant="select-light"
-        state="selected"
+      <Interactive.Base
+        variant="select"
+        prominence="light"
+        selected
         onClick={() => {}}
       >
         <Interactive.Container border>
-          <span className="interactive-foreground">Selected (light)</span>
+          <span>Selected (light)</span>
         </Interactive.Container>
-      </Interactive.Stateful>
+      </Interactive.Base>
 
-      <Interactive.Stateful
-        variant="select-light"
-        state="empty"
-        onClick={() => {}}
-      >
+      <Interactive.Base variant="select" prominence="light" onClick={() => {}}>
         <Interactive.Container border>
-          <span className="interactive-foreground">Unselected (light)</span>
+          <span>Unselected (light)</span>
         </Interactive.Container>
-      </Interactive.Stateful>
+      </Interactive.Base>
 
-      <Interactive.Stateful
-        variant="select-heavy"
-        state="selected"
+      <Interactive.Base
+        variant="select"
+        prominence="heavy"
+        selected
         onClick={() => {}}
       >
         <Interactive.Container border>
-          <span className="interactive-foreground">Selected (heavy)</span>
+          <span>Selected (heavy)</span>
         </Interactive.Container>
-      </Interactive.Stateful>
+      </Interactive.Base>
 
-      <Interactive.Stateful
-        variant="select-heavy"
-        state="empty"
-        onClick={() => {}}
-      >
+      <Interactive.Base variant="select" prominence="heavy" onClick={() => {}}>
         <Interactive.Container border>
-          <span className="interactive-foreground">Unselected (heavy)</span>
+          <span>Unselected (heavy)</span>
         </Interactive.Container>
-      </Interactive.Stateful>
+      </Interactive.Base>
     </div>
   ),
 };

web/lib/opal/src/icons/index.ts

@@ -89,7 +89,7 @@ export { default as SvgHistory } from "@opal/icons/history";
 export { default as SvgHourglass } from "@opal/icons/hourglass";
 export { default as SvgImage } from "@opal/icons/image";
 export { default as SvgImageSmall } from "@opal/icons/image-small";
-export { default as SvgImport } from "@opal/icons/import-icon";
+export { default as SvgImport } from "@opal/icons/import";
 export { default as SvgInfo } from "@opal/icons/info";
 export { default as SvgInfoSmall } from "@opal/icons/info-small";
 export { default as SvgKey } from "@opal/icons/key";

web/lib/opal/src/layouts/Content/BodyLayout.stories.tsx

@@ -0,0 +1,87 @@
+import type { Meta, StoryObj } from "@storybook/react";
+import { BodyLayout } from "./BodyLayout";
+import { SvgSettings, SvgStar, SvgRefreshCw } from "@opal/icons";
+
+const meta = {
+  title: "Layouts/BodyLayout",
+  component: BodyLayout,
+  tags: ["autodocs"],
+  parameters: {
+    layout: "centered",
+  },
+} satisfies Meta<typeof BodyLayout>;
+
+export default meta;
+
+type Story = StoryObj<typeof meta>;
+
+// ---------------------------------------------------------------------------
+// Size presets
+// ---------------------------------------------------------------------------
+
+export const MainContent: Story = {
+  args: {
+    sizePreset: "main-content",
+    title: "Last synced 2 minutes ago",
+  },
+};
+
+export const MainUi: Story = {
+  args: {
+    sizePreset: "main-ui",
+    title: "Document count: 1,234",
+  },
+};
+
+export const Secondary: Story = {
+  args: {
+    sizePreset: "secondary",
+    title: "Updated 5 min ago",
+  },
+};
+
+// ---------------------------------------------------------------------------
+// With icon
+// ---------------------------------------------------------------------------
+
+export const WithIcon: Story = {
+  args: {
+    sizePreset: "main-ui",
+    title: "Settings",
+    icon: SvgSettings,
+  },
+};
+
+// ---------------------------------------------------------------------------
+// Orientations
+// ---------------------------------------------------------------------------
+
+export const Vertical: Story = {
+  args: {
+    sizePreset: "main-ui",
+    title: "Stacked layout",
+    icon: SvgStar,
+    orientation: "vertical",
+  },
+};
+
+export const Reverse: Story = {
+  args: {
+    sizePreset: "main-ui",
+    title: "Reverse layout",
+    icon: SvgRefreshCw,
+    orientation: "reverse",
+  },
+};
+
+// ---------------------------------------------------------------------------
+// Prominence
+// ---------------------------------------------------------------------------
+
+export const Muted: Story = {
+  args: {
+    sizePreset: "main-ui",
+    title: "Muted body text",
+    prominence: "muted",
+  },
+};

web/lib/opal/src/layouts/Content/HeadingLayout.stories.tsx

@@ -0,0 +1,98 @@
+import type { Meta, StoryObj } from "@storybook/react";
+import { HeadingLayout } from "./HeadingLayout";
+import { SvgSettings, SvgStar } from "@opal/icons";
+import * as TooltipPrimitive from "@radix-ui/react-tooltip";
+
+const meta = {
+  title: "Layouts/HeadingLayout",
+  component: HeadingLayout,
+  tags: ["autodocs"],
+  parameters: {
+    layout: "centered",
+  },
+  decorators: [
+    (Story) => (
+      <TooltipPrimitive.Provider>
+        <Story />
+      </TooltipPrimitive.Provider>
+    ),
+  ],
+} satisfies Meta<typeof HeadingLayout>;
+
+export default meta;
+
+type Story = StoryObj<typeof meta>;
+
+// ---------------------------------------------------------------------------
+// Size presets
+// ---------------------------------------------------------------------------
+
+export const Headline: Story = {
+  args: {
+    sizePreset: "headline",
+    title: "Welcome to Onyx",
+    description: "Your enterprise search and AI assistant platform.",
+  },
+};
+
+export const Section: Story = {
+  args: {
+    sizePreset: "section",
+    title: "Configuration",
+  },
+};
+
+// ---------------------------------------------------------------------------
+// With icon
+// ---------------------------------------------------------------------------
+
+export const WithIcon: Story = {
+  args: {
+    sizePreset: "headline",
+    title: "Settings",
+    icon: SvgSettings,
+  },
+};
+
+export const SectionWithIcon: Story = {
+  args: {
+    sizePreset: "section",
+    variant: "section",
+    title: "Favorites",
+    icon: SvgStar,
+  },
+};
+
+// ---------------------------------------------------------------------------
+// Variants
+// ---------------------------------------------------------------------------
+
+export const SectionVariant: Story = {
+  args: {
+    sizePreset: "headline",
+    variant: "section",
+    title: "Inline Icon Heading",
+    icon: SvgSettings,
+  },
+};
+
+// ---------------------------------------------------------------------------
+// Editable
+// ---------------------------------------------------------------------------
+
+export const Editable: Story = {
+  args: {
+    sizePreset: "headline",
+    title: "Click to edit me",
+    editable: true,
+  },
+};
+
+export const EditableSection: Story = {
+  args: {
+    sizePreset: "section",
+    title: "Editable Section Title",
+    editable: true,
+    description: "This title can be edited inline.",
+  },
+};

web/lib/opal/src/layouts/Content/LabelLayout.stories.tsx

@@ -0,0 +1,154 @@
+import type { Meta, StoryObj } from "@storybook/react";
+import { LabelLayout } from "./LabelLayout";
+import { SvgSettings, SvgStar } from "@opal/icons";
+import * as TooltipPrimitive from "@radix-ui/react-tooltip";
+
+const meta = {
+  title: "Layouts/LabelLayout",
+  component: LabelLayout,
+  tags: ["autodocs"],
+  parameters: {
+    layout: "centered",
+  },
+  decorators: [
+    (Story) => (
+      <TooltipPrimitive.Provider>
+        <Story />
+      </TooltipPrimitive.Provider>
+    ),
+  ],
+} satisfies Meta<typeof LabelLayout>;
+
+export default meta;
+
+type Story = StoryObj<typeof meta>;
+
+// ---------------------------------------------------------------------------
+// Size presets
+// ---------------------------------------------------------------------------
+
+export const MainContent: Story = {
+  args: {
+    sizePreset: "main-content",
+    title: "Display Name",
+  },
+};
+
+export const MainUi: Story = {
+  args: {
+    sizePreset: "main-ui",
+    title: "Email Address",
+  },
+};
+
+export const SecondaryPreset: Story = {
+  args: {
+    sizePreset: "secondary",
+    title: "API Key",
+  },
+};
+
+// ---------------------------------------------------------------------------
+// With description
+// ---------------------------------------------------------------------------
+
+export const WithDescription: Story = {
+  args: {
+    sizePreset: "main-content",
+    title: "Workspace Name",
+    description: "The name displayed across your organization.",
+  },
+};
+
+// ---------------------------------------------------------------------------
+// With icon
+// ---------------------------------------------------------------------------
+
+export const WithIcon: Story = {
+  args: {
+    sizePreset: "main-ui",
+    title: "Settings",
+    icon: SvgSettings,
+  },
+};
+
+// ---------------------------------------------------------------------------
+// Optional
+// ---------------------------------------------------------------------------
+
+export const Optional: Story = {
+  args: {
+    sizePreset: "main-content",
+    title: "Phone Number",
+    optional: true,
+  },
+};
+
+// ---------------------------------------------------------------------------
+// Aux icons
+// ---------------------------------------------------------------------------
+
+export const AuxInfoGray: Story = {
+  args: {
+    sizePreset: "main-content",
+    title: "Connection Status",
+    auxIcon: "info-gray",
+  },
+};
+
+export const AuxWarning: Story = {
+  args: {
+    sizePreset: "main-content",
+    title: "Rate Limit",
+    auxIcon: "warning",
+  },
+};
+
+export const AuxError: Story = {
+  args: {
+    sizePreset: "main-content",
+    title: "API Key",
+    auxIcon: "error",
+  },
+};
+
+// ---------------------------------------------------------------------------
+// With tag
+// ---------------------------------------------------------------------------
+
+export const WithTag: Story = {
+  args: {
+    sizePreset: "main-ui",
+    title: "Knowledge Graph",
+    tag: { title: "Beta", color: "blue" },
+  },
+};
+
+// ---------------------------------------------------------------------------
+// Editable
+// ---------------------------------------------------------------------------
+
+export const Editable: Story = {
+  args: {
+    sizePreset: "main-ui",
+    title: "Click to edit",
+    editable: true,
+  },
+};
+
+// ---------------------------------------------------------------------------
+// Combined
+// ---------------------------------------------------------------------------
+
+export const FullFeatured: Story = {
+  args: {
+    sizePreset: "main-content",
+    title: "Custom Field",
+    icon: SvgStar,
+    description: "A custom field with all extras enabled.",
+    optional: true,
+    auxIcon: "info-blue",
+    tag: { title: "New", color: "green" },
+    editable: true,
+  },
+};

web/lib/opal/src/layouts/content/BodyLayout.tsx

@@ -0,0 +1,134 @@
+"use client";
+
+import type { IconFunctionComponent } from "@opal/types";
+import { cn } from "@opal/utils";
+
+// ---------------------------------------------------------------------------
+// Types
+// ---------------------------------------------------------------------------
+
+type BodySizePreset = "main-content" | "main-ui" | "secondary";
+type BodyOrientation = "vertical" | "inline" | "reverse";
+type BodyProminence = "default" | "muted";
+
+interface BodyPresetConfig {
+  /** Icon width/height (CSS value). */
+  iconSize: string;
+  /** Tailwind padding class for the icon container. */
+  iconContainerPadding: string;
+  /** Tailwind font class for the title. */
+  titleFont: string;
+  /** Title line-height — also used as icon container min-height (CSS value). */
+  lineHeight: string;
+  /** Gap between icon container and title (CSS value). */
+  gap: string;
+}
+
+/** Props for {@link BodyLayout}. Does not support editing or descriptions. */
+interface BodyLayoutProps {
+  /** Optional icon component. */
+  icon?: IconFunctionComponent;
+
+  /** Main title text (read-only — editing is not supported). */
+  title: string;
+
+  /** Size preset. Default: `"main-ui"`. */
+  sizePreset?: BodySizePreset;
+
+  /** Layout orientation. Default: `"inline"`. */
+  orientation?: BodyOrientation;
+
+  /** Title prominence. Default: `"default"`. */
+  prominence?: BodyProminence;
+
+  /** Ref forwarded to the root `<div>`. */
+  ref?: React.Ref<HTMLDivElement>;
+}
+
+// ---------------------------------------------------------------------------
+// Presets
+// ---------------------------------------------------------------------------
+
+const BODY_PRESETS: Record<BodySizePreset, BodyPresetConfig> = {
+  "main-content": {
+    iconSize: "1rem",
+    iconContainerPadding: "p-1",
+    titleFont: "font-main-content-body",
+    lineHeight: "1.5rem",
+    gap: "0.125rem",
+  },
+  "main-ui": {
+    iconSize: "1rem",
+    iconContainerPadding: "p-0.5",
+    titleFont: "font-main-ui-action",
+    lineHeight: "1.25rem",
+    gap: "0.25rem",
+  },
+  secondary: {
+    iconSize: "0.75rem",
+    iconContainerPadding: "p-0.5",
+    titleFont: "font-secondary-action",
+    lineHeight: "1rem",
+    gap: "0.125rem",
+  },
+};
+
+// ---------------------------------------------------------------------------
+// BodyLayout
+// ---------------------------------------------------------------------------
+
+function BodyLayout({
+  icon: Icon,
+  title,
+  sizePreset = "main-ui",
+  orientation = "inline",
+  prominence = "default",
+  ref,
+}: BodyLayoutProps) {
+  const config = BODY_PRESETS[sizePreset];
+  const titleColorClass =
+    prominence === "muted" ? "text-text-03" : "text-text-04";
+
+  return (
+    <div
+      ref={ref}
+      className="opal-content-body"
+      data-orientation={orientation}
+      style={{ gap: config.gap }}
+    >
+      {Icon && (
+        <div
+          className={cn(
+            "opal-content-body-icon-container shrink-0",
+            config.iconContainerPadding
+          )}
+          style={{ minHeight: config.lineHeight }}
+        >
+          <Icon
+            className="opal-content-body-icon text-text-03"
+            style={{ width: config.iconSize, height: config.iconSize }}
+          />
+        </div>
+      )}
+
+      <span
+        className={cn(
+          "opal-content-body-title",
+          config.titleFont,
+          titleColorClass
+        )}
+        style={{ height: config.lineHeight }}
+      >
+        {title}
+      </span>
+    </div>
+  );
+}
+
+export {
+  BodyLayout,
+  type BodyLayoutProps,
+  type BodySizePreset,
+  type BodyOrientation,
+  type BodyProminence,
+};

web/lib/opal/src/layouts/content/HeadingLayout.tsx

@@ -0,0 +1,218 @@
+"use client";
+
+import { Button } from "@opal/components/buttons/button/components";
+import type { SizeVariant } from "@opal/shared";
+import SvgEdit from "@opal/icons/edit";
+import type { IconFunctionComponent } from "@opal/types";
+import { cn } from "@opal/utils";
+import { useRef, useState } from "react";
+
+// ---------------------------------------------------------------------------
+// Types
+// ---------------------------------------------------------------------------
+
+type HeadingSizePreset = "headline" | "section";
+type HeadingVariant = "heading" | "section";
+
+interface HeadingPresetConfig {
+  /** Icon width/height (CSS value). */
+  iconSize: string;
+  /** Tailwind padding class for the icon container. */
+  iconContainerPadding: string;
+  /** Gap between icon container and content (CSS value). */
+  gap: string;
+  /** Tailwind font class for the title. */
+  titleFont: string;
+  /** Title line-height — also used as icon container min-height (CSS value). */
+  lineHeight: string;
+  /** Button `size` prop for the edit button. Uses the shared `SizeVariant` scale. */
+  editButtonSize: SizeVariant;
+  /** Tailwind padding class for the edit button container. */
+  editButtonPadding: string;
+}
+
+interface HeadingLayoutProps {
+  /** Optional icon component. */
+  icon?: IconFunctionComponent;
+
+  /** Main title text. */
+  title: string;
+
+  /** Optional description below the title. */
+  description?: string;
+
+  /** Enable inline editing of the title. */
+  editable?: boolean;
+
+  /** Called when the user commits an edit. */
+  onTitleChange?: (newTitle: string) => void;
+
+  /** Size preset. Default: `"headline"`. */
+  sizePreset?: HeadingSizePreset;
+
+  /** Variant controls icon placement. `"heading"` = top, `"section"` = inline. Default: `"heading"`. */
+  variant?: HeadingVariant;
+
+  /** Ref forwarded to the root `<div>`. */
+  ref?: React.Ref<HTMLDivElement>;
+}
+
+// ---------------------------------------------------------------------------
+// Presets
+// ---------------------------------------------------------------------------
+
+const HEADING_PRESETS: Record<HeadingSizePreset, HeadingPresetConfig> = {
+  headline: {
+    iconSize: "2rem",
+    iconContainerPadding: "p-0.5",
+    gap: "0.25rem",
+    titleFont: "font-heading-h2",
+    lineHeight: "2.25rem",
+    editButtonSize: "md",
+    editButtonPadding: "p-1",
+  },
+  section: {
+    iconSize: "1.25rem",
+    iconContainerPadding: "p-1",
+    gap: "0rem",
+    titleFont: "font-heading-h3",
+    lineHeight: "1.75rem",
+    editButtonSize: "sm",
+    editButtonPadding: "p-0.5",
+  },
+};
+
+// ---------------------------------------------------------------------------
+// HeadingLayout
+// ---------------------------------------------------------------------------
+
+function HeadingLayout({
+  sizePreset = "headline",
+  variant = "heading",
+  icon: Icon,
+  title,
+  description,
+  editable,
+  onTitleChange,
+  ref,
+}: HeadingLayoutProps) {
+  const [editing, setEditing] = useState(false);
+  const [editValue, setEditValue] = useState(title);
+  const inputRef = useRef<HTMLInputElement>(null);
+
+  const config = HEADING_PRESETS[sizePreset];
+  const iconPlacement = variant === "heading" ? "top" : "left";
+
+  function startEditing() {
+    setEditValue(title);
+    setEditing(true);
+  }
+
+  function commit() {
+    const value = editValue.trim();
+    if (value && value !== title) onTitleChange?.(value);
+    setEditing(false);
+  }
+
+  return (
+    <div
+      ref={ref}
+      className="opal-content-heading"
+      data-icon-placement={iconPlacement}
+      style={{ gap: iconPlacement === "left" ? config.gap : undefined }}
+    >
+      {Icon && (
+        <div
+          className={cn(
+            "opal-content-heading-icon-container shrink-0",
+            config.iconContainerPadding
+          )}
+          style={{ minHeight: config.lineHeight }}
+        >
+          <Icon
+            className="opal-content-heading-icon"
+            style={{ width: config.iconSize, height: config.iconSize }}
+          />
+        </div>
+      )}
+
+      <div className="opal-content-heading-body">
+        <div className="opal-content-heading-title-row">
+          {editing ? (
+            <div className="opal-content-heading-input-sizer">
+              <span
+                className={cn(
+                  "opal-content-heading-input-mirror",
+                  config.titleFont
+                )}
+              >
+                {editValue || "\u00A0"}
+              </span>
+              <input
+                ref={inputRef}
+                className={cn(
+                  "opal-content-heading-input",
+                  config.titleFont,
+                  "text-text-04"
+                )}
+                value={editValue}
+                onChange={(e) => setEditValue(e.target.value)}
+                size={1}
+                autoFocus
+                onFocus={(e) => e.currentTarget.select()}
+                onBlur={commit}
+                onKeyDown={(e) => {
+                  if (e.key === "Enter") commit();
+                  if (e.key === "Escape") {
+                    setEditValue(title);
+                    setEditing(false);
+                  }
+                }}
+                style={{ height: config.lineHeight }}
+              />
+            </div>
+          ) : (
+            <span
+              className={cn(
+                "opal-content-heading-title",
+                config.titleFont,
+                "text-text-04",
+                editable && "cursor-pointer"
+              )}
+              onClick={editable ? startEditing : undefined}
+              style={{ height: config.lineHeight }}
+            >
+              {title}
+            </span>
+          )}
+
+          {editable && !editing && (
+            <div
+              className={cn(
+                "opal-content-heading-edit-button",
+                config.editButtonPadding
+              )}
+            >
+              <Button
+                icon={SvgEdit}
+                prominence="internal"
+                size={config.editButtonSize}
+                tooltip="Edit"
+                tooltipSide="right"
+                onClick={startEditing}
+              />
+            </div>
+          )}
+        </div>
+
+        {description && (
+          <div className="opal-content-heading-description font-secondary-body text-text-03">
+            {description}
+          </div>
+        )}
+      </div>
+    </div>
+  );
+}
+
+export { HeadingLayout, type HeadingLayoutProps, type HeadingSizePreset };

web/lib/opal/src/layouts/content/LabelLayout.tsx

@@ -0,0 +1,286 @@
+"use client";
+
+import { Button } from "@opal/components/buttons/button/components";
+import { Tag, type TagProps } from "@opal/components/tag/components";
+import type { SizeVariant } from "@opal/shared";
+import SvgAlertCircle from "@opal/icons/alert-circle";
+import SvgAlertTriangle from "@opal/icons/alert-triangle";
+import SvgEdit from "@opal/icons/edit";
+import SvgXOctagon from "@opal/icons/x-octagon";
+import type { IconFunctionComponent } from "@opal/types";
+import { cn } from "@opal/utils";
+import { useRef, useState } from "react";
+
+// ---------------------------------------------------------------------------
+// Types
+// ---------------------------------------------------------------------------
+
+type LabelSizePreset = "main-content" | "main-ui" | "secondary";
+
+type LabelAuxIcon = "info-gray" | "info-blue" | "warning" | "error";
+
+interface LabelPresetConfig {
+  iconSize: string;
+  iconContainerPadding: string;
+  iconColorClass: string;
+  titleFont: string;
+  lineHeight: string;
+  gap: string;
+  /** Button `size` prop for the edit button. Uses the shared `SizeVariant` scale. */
+  editButtonSize: SizeVariant;
+  editButtonPadding: string;
+  optionalFont: string;
+  /** Aux icon size = lineHeight − 2 × p-0.5. */
+  auxIconSize: string;
+}
+
+interface LabelLayoutProps {
+  /** Optional icon component. */
+  icon?: IconFunctionComponent;
+
+  /** Main title text. */
+  title: string;
+
+  /** Optional description text below the title. */
+  description?: string;
+
+  /** Enable inline editing of the title. */
+  editable?: boolean;
+
+  /** Called when the user commits an edit. */
+  onTitleChange?: (newTitle: string) => void;
+
+  /** When `true`, renders "(Optional)" beside the title. */
+  optional?: boolean;
+
+  /** Auxiliary status icon rendered beside the title. */
+  auxIcon?: LabelAuxIcon;
+
+  /** Tag rendered beside the title. */
+  tag?: TagProps;
+
+  /** Size preset. Default: `"main-ui"`. */
+  sizePreset?: LabelSizePreset;
+
+  /** Ref forwarded to the root `<div>`. */
+  ref?: React.Ref<HTMLDivElement>;
+}
+
+// ---------------------------------------------------------------------------
+// Presets
+// ---------------------------------------------------------------------------
+
+const LABEL_PRESETS: Record<LabelSizePreset, LabelPresetConfig> = {
+  "main-content": {
+    iconSize: "1rem",
+    iconContainerPadding: "p-1",
+    iconColorClass: "text-text-04",
+    titleFont: "font-main-content-emphasis",
+    lineHeight: "1.5rem",
+    gap: "0.125rem",
+    editButtonSize: "sm",
+    editButtonPadding: "p-0",
+    optionalFont: "font-main-content-muted",
+    auxIconSize: "1.25rem",
+  },
+  "main-ui": {
+    iconSize: "1rem",
+    iconContainerPadding: "p-0.5",
+    iconColorClass: "text-text-03",
+    titleFont: "font-main-ui-action",
+    lineHeight: "1.25rem",
+    gap: "0.25rem",
+    editButtonSize: "xs",
+    editButtonPadding: "p-0",
+    optionalFont: "font-main-ui-muted",
+    auxIconSize: "1rem",
+  },
+  secondary: {
+    iconSize: "0.75rem",
+    iconContainerPadding: "p-0.5",
+    iconColorClass: "text-text-04",
+    titleFont: "font-secondary-action",
+    lineHeight: "1rem",
+    gap: "0.125rem",
+    editButtonSize: "2xs",
+    editButtonPadding: "p-0",
+    optionalFont: "font-secondary-action",
+    auxIconSize: "0.75rem",
+  },
+};
+
+// ---------------------------------------------------------------------------
+// LabelLayout
+// ---------------------------------------------------------------------------
+
+const AUX_ICON_CONFIG: Record<
+  LabelAuxIcon,
+  { icon: IconFunctionComponent; colorClass: string }
+> = {
+  "info-gray": { icon: SvgAlertCircle, colorClass: "text-text-02" },
+  "info-blue": { icon: SvgAlertCircle, colorClass: "text-status-info-05" },
+  warning: { icon: SvgAlertTriangle, colorClass: "text-status-warning-05" },
+  error: { icon: SvgXOctagon, colorClass: "text-status-error-05" },
+};
+
+function LabelLayout({
+  icon: Icon,
+  title,
+  description,
+  editable,
+  onTitleChange,
+  optional,
+  auxIcon,
+  tag,
+  sizePreset = "main-ui",
+  ref,
+}: LabelLayoutProps) {
+  const [editing, setEditing] = useState(false);
+  const [editValue, setEditValue] = useState(title);
+  const inputRef = useRef<HTMLInputElement>(null);
+
+  const config = LABEL_PRESETS[sizePreset];
+
+  function startEditing() {
+    setEditValue(title);
+    setEditing(true);
+  }
+
+  function commit() {
+    const value = editValue.trim();
+    if (value && value !== title) onTitleChange?.(value);
+    setEditing(false);
+  }
+
+  return (
+    <div ref={ref} className="opal-content-label" style={{ gap: config.gap }}>
+      {Icon && (
+        <div
+          className={cn(
+            "opal-content-label-icon-container shrink-0",
+            config.iconContainerPadding
+          )}
+          style={{ minHeight: config.lineHeight }}
+        >
+          <Icon
+            className={cn("opal-content-label-icon", config.iconColorClass)}
+            style={{ width: config.iconSize, height: config.iconSize }}
+          />
+        </div>
+      )}
+
+      <div className="opal-content-label-body">
+        <div className="opal-content-label-title-row">
+          {editing ? (
+            <div className="opal-content-label-input-sizer">
+              <span
+                className={cn(
+                  "opal-content-label-input-mirror",
+                  config.titleFont
+                )}
+              >
+                {editValue || "\u00A0"}
+              </span>
+              <input
+                ref={inputRef}
+                className={cn(
+                  "opal-content-label-input",
+                  config.titleFont,
+                  "text-text-04"
+                )}
+                value={editValue}
+                onChange={(e) => setEditValue(e.target.value)}
+                size={1}
+                autoFocus
+                onFocus={(e) => e.currentTarget.select()}
+                onBlur={commit}
+                onKeyDown={(e) => {
+                  if (e.key === "Enter") commit();
+                  if (e.key === "Escape") {
+                    setEditValue(title);
+                    setEditing(false);
+                  }
+                }}
+                style={{ height: config.lineHeight }}
+              />
+            </div>
+          ) : (
+            <span
+              className={cn(
+                "opal-content-label-title",
+                config.titleFont,
+                "text-text-04",
+                editable && "cursor-pointer"
+              )}
+              onClick={editable ? startEditing : undefined}
+              style={{ height: config.lineHeight }}
+            >
+              {title}
+            </span>
+          )}
+
+          {optional && (
+            <span
+              className={cn(config.optionalFont, "text-text-03 shrink-0")}
+              style={{ height: config.lineHeight }}
+            >
+              (Optional)
+            </span>
+          )}
+
+          {auxIcon &&
+            (() => {
+              const { icon: AuxIcon, colorClass } = AUX_ICON_CONFIG[auxIcon];
+              return (
+                <div
+                  className="opal-content-label-aux-icon shrink-0 p-0.5"
+                  style={{ height: config.lineHeight }}
+                >
+                  <AuxIcon
+                    className={colorClass}
+                    style={{
+                      width: config.auxIconSize,
+                      height: config.auxIconSize,
+                    }}
+                  />
+                </div>
+              );
+            })()}
+
+          {tag && <Tag {...tag} />}
+
+          {editable && !editing && (
+            <div
+              className={cn(
+                "opal-content-label-edit-button",
+                config.editButtonPadding
+              )}
+            >
+              <Button
+                icon={SvgEdit}
+                prominence="internal"
+                size={config.editButtonSize}
+                tooltip="Edit"
+                tooltipSide="right"
+                onClick={startEditing}
+              />
+            </div>
+          )}
+        </div>
+
+        {description && (
+          <div className="opal-content-label-description font-secondary-body text-text-03">
+            {description}
+          </div>
+        )}
+      </div>
+    </div>
+  );
+}
+
+export {
+  LabelLayout,
+  type LabelLayoutProps,
+  type LabelSizePreset,
+  type LabelAuxIcon,
+};

web/src/app/admin/configuration/llm/utils.ts

@@ -14,7 +14,6 @@ import {
   QwenIcon,
   OllamaIcon,
   LMStudioIcon,
-  LiteLLMIcon,
   ZAIIcon,
 } from "@/components/icons/icons";
 import {
@@ -22,14 +21,12 @@ import {
   OpenRouterModelResponse,
   BedrockModelResponse,
   LMStudioModelResponse,
-  LiteLLMProxyModelResponse,
   ModelConfiguration,
   LLMProviderName,
   BedrockFetchParams,
   OllamaFetchParams,
   LMStudioFetchParams,
   OpenRouterFetchParams,
-  LiteLLMProxyFetchParams,
 } from "@/interfaces/llm";
 import { SvgAws, SvgOpenrouter } from "@opal/icons";
 
@@ -40,7 +37,6 @@ export const AGGREGATOR_PROVIDERS = new Set([
   "openrouter",
   "ollama_chat",
   "lm_studio",
-  "litellm_proxy",
   "vertex_ai",
 ]);
 
@@ -77,7 +73,6 @@ export const getProviderIcon = (
     bedrock: SvgAws,
     bedrock_converse: SvgAws,
     openrouter: SvgOpenrouter,
-    litellm_proxy: LiteLLMIcon,
     vertex_ai: GeminiIcon,
   };
 
@@ -343,65 +338,6 @@ export const fetchLMStudioModels = async (
   }
 };
 
-/**
- * Fetches LiteLLM Proxy models directly without any form state dependencies.
- * Uses snake_case params to match API structure.
- */
-export const fetchLiteLLMProxyModels = async (
-  params: LiteLLMProxyFetchParams
-): Promise<{ models: ModelConfiguration[]; error?: string }> => {
-  const apiBase = params.api_base;
-  const apiKey = params.api_key;
-  if (!apiBase) {
-    return { models: [], error: "API Base is required" };
-  }
-  if (!apiKey) {
-    return { models: [], error: "API Key is required" };
-  }
-
-  try {
-    const response = await fetch("/api/admin/llm/litellm/available-models", {
-      method: "POST",
-      headers: {
-        "Content-Type": "application/json",
-      },
-      body: JSON.stringify({
-        api_base: apiBase,
-        api_key: apiKey,
-        provider_name: params.provider_name,
-      }),
-      signal: params.signal,
-    });
-
-    if (!response.ok) {
-      let errorMessage = "Failed to fetch models";
-      try {
-        const errorData = await response.json();
-        errorMessage = errorData.detail || errorData.message || errorMessage;
-      } catch {
-        // ignore JSON parsing errors
-      }
-      return { models: [], error: errorMessage };
-    }
-
-    const data: LiteLLMProxyModelResponse[] = await response.json();
-    const models: ModelConfiguration[] = data.map((modelData) => ({
-      name: modelData.model_name,
-      display_name: modelData.model_name,
-      is_visible: true,
-      max_input_tokens: null,
-      supports_image_input: false,
-      supports_reasoning: false,
-    }));
-
-    return { models };
-  } catch (error) {
-    const errorMessage =
-      error instanceof Error ? error.message : "Unknown error";
-    return { models: [], error: errorMessage };
-  }
-};
-
 /**
  * Fetches models for a provider. Accepts form values directly and maps them
  * to the expected fetch params format internally.
@@ -449,13 +385,6 @@ export const fetchModels = async (
         api_key: formValues.api_key,
         provider_name: formValues.name,
       });
-    case LLMProviderName.LITELLM_PROXY:
-      return fetchLiteLLMProxyModels({
-        api_base: formValues.api_base,
-        api_key: formValues.api_key,
-        provider_name: formValues.name,
-        signal,
-      });
     default:
       return { models: [], error: `Unknown provider: ${providerName}` };
   }
@@ -468,7 +397,6 @@ export function canProviderFetchModels(providerName?: string) {
     case LLMProviderName.OLLAMA_CHAT:
     case LLMProviderName.LM_STUDIO:
     case LLMProviderName.OPENROUTER:
-    case LLMProviderName.LITELLM_PROXY:
       return true;
     default:
       return false;

web/src/app/admin/users2/page.tsx

@@ -1 +0,0 @@
-export { default } from "@/refresh-pages/admin/UsersPage";

web/src/app/app/message/messageComponents/MessageToolbar.tsx

@@ -249,7 +249,6 @@ export default function MessageToolbar({
             <SelectButton
               icon={SvgThumbsUp}
               onClick={() => handleFeedbackClick("like")}
-              variant="select-light"
               state={isFeedbackTransient("like") ? "selected" : "empty"}
               tooltip={
                 currentFeedback === "like" ? "Remove Like" : "Good Response"
@@ -259,7 +258,6 @@ export default function MessageToolbar({
             <SelectButton
               icon={SvgThumbsDown}
               onClick={() => handleFeedbackClick("dislike")}
-              variant="select-light"
               state={isFeedbackTransient("dislike") ? "selected" : "empty"}
               tooltip={
                 currentFeedback === "dislike"
@@ -285,7 +283,7 @@ export default function MessageToolbar({
                       });
                       regenerator(llmDescriptor);
                     }}
-                    foldable
+                    folded
                   />
                 </div>
               )}

web/src/app/craft/components/output-panel/ArtifactsTab.tsx

@@ -98,7 +98,7 @@ export default function ArtifactsTab({
   const handleWebappDownload = () => {
     if (!sessionId) return;
     const link = document.createElement("a");
-    link.href = `/api/build/sessions/${sessionId}/webapp-download`;
+    link.href = `/api/build/sessions/${sessionId}/webapp/download`;
     link.download = "";
     document.body.appendChild(link);
     link.click();

web/src/app/nrf/NRFChrome.tsx

@@ -11,7 +11,7 @@ import { Button } from "@opal/components";
 import { SvgBubbleText, SvgSearchMenu, SvgSidebar } from "@opal/icons";
 import MinimalMarkdown from "@/components/chat/MinimalMarkdown";
 import { useSettingsContext } from "@/providers/SettingsProvider";
-import type { AppMode } from "@/providers/QueryControllerProvider";
+import { AppMode, useAppMode } from "@/providers/AppModeProvider";
 import useAppFocus from "@/hooks/useAppFocus";
 import { useQueryController } from "@/providers/QueryControllerProvider";
 import { usePaidEnterpriseFeaturesEnabled } from "@/components/settings/usePaidEnterpriseFeaturesEnabled";
@@ -58,15 +58,15 @@ const footerMarkdownComponents = {
  */
 export default function NRFChrome() {
   const isPaidEnterpriseFeaturesEnabled = usePaidEnterpriseFeaturesEnabled();
-  const { state, setAppMode } = useQueryController();
+  const { appMode, setAppMode } = useAppMode();
   const settings = useSettingsContext();
   const { isMobile } = useScreenSize();
   const { setFolded } = useAppSidebarContext();
   const appFocus = useAppFocus();
+  const { classification } = useQueryController();
   const [modePopoverOpen, setModePopoverOpen] = useState(false);
 
-  const effectiveMode: AppMode =
-    appFocus.isNewSession() && state.phase === "idle" ? state.appMode : "chat";
+  const effectiveMode: AppMode = appFocus.isNewSession() ? appMode : "chat";
 
   const customFooterContent =
     settings?.enterpriseSettings?.custom_lower_disclaimer_content ||
@@ -78,7 +78,7 @@ export default function NRFChrome() {
     isPaidEnterpriseFeaturesEnabled &&
     settings.isSearchModeAvailable &&
     appFocus.isNewSession() &&
-    state.phase === "idle";
+    !classification;
 
   const showHeader = isMobile || showModeToggle;
 

web/src/app/nrf/NRFPage.tsx

@@ -175,7 +175,7 @@ export default function NRFPage({ isSidePanel = false }: NRFPageProps) {
   const isStreaming = currentChatState === "streaming";
 
   // Query controller for search/chat classification (EE feature)
-  const { submit: submitQuery, state } = useQueryController();
+  const { submit: submitQuery, classification } = useQueryController();
 
   // Determine if retrieval (search) is enabled based on the agent
   const retrievalEnabled = useMemo(() => {
@@ -186,8 +186,7 @@ export default function NRFPage({ isSidePanel = false }: NRFPageProps) {
   }, [liveAgent]);
 
   // Check if we're in search mode
-  const isSearch =
-    state.phase === "searching" || state.phase === "search-results";
+  const isSearch = classification === "search";
 
   // Anchor for scroll positioning (matches ChatPage pattern)
   const anchorMessage = messageHistory.at(-2) ?? messageHistory[0];
@@ -318,7 +317,7 @@ export default function NRFPage({ isSidePanel = false }: NRFPageProps) {
       };
 
       // Use submitQuery which will classify the query and either:
-      // - Route to search (sets phase to "searching"/"search-results" and shows SearchUI)
+      // - Route to search (sets classification to "search" and shows SearchUI)
       // - Route to chat (calls onChat callback)
       await submitQuery(submittedMessage, onChat);
     },

web/src/components/admin/ClientLayout.tsx

@@ -31,7 +31,6 @@ const SETTINGS_LAYOUT_PREFIXES = [
   ADMIN_PATHS.LLM_MODELS,
   ADMIN_PATHS.AGENTS,
   ADMIN_PATHS.USERS,
-  ADMIN_PATHS.USERS_V2,
   ADMIN_PATHS.TOKEN_RATE_LIMITS,
   ADMIN_PATHS.SEARCH_SETTINGS,
   ADMIN_PATHS.DOCUMENT_PROCESSING,

web/src/components/tools/CSVContent.tsx

@@ -60,28 +60,27 @@ const CsvContent: React.FC<ContentComponentProps> = ({
       }
 
       const csvData = await response.text();
-      const rows = parseCSV(csvData.trim());
+      const rows = csvData.trim().split("\n");
       const firstRow = rows[0];
       if (!firstRow) {
         throw new Error("CSV file is empty");
       }
-      const parsedHeaders = firstRow;
+      const parsedHeaders = firstRow.split(",");
       setHeaders(parsedHeaders);
 
-      const parsedData: Record<string, string>[] = rows
-        .slice(1)
-        .map((fields) => {
-          return parsedHeaders.reduce<Record<string, string>>(
-            (obj, header, index) => {
-              const val = fields[index];
-              if (val !== undefined) {
-                obj[header] = val;
-              }
-              return obj;
-            },
-            {}
-          );
-        });
+      const parsedData: Record<string, string>[] = rows.slice(1).map((row) => {
+        const values = row.split(",");
+        return parsedHeaders.reduce<Record<string, string>>(
+          (obj, header, index) => {
+            const val = values[index];
+            if (val !== undefined) {
+              obj[header] = val;
+            }
+            return obj;
+          },
+          {}
+        );
+      });
       setData(parsedData);
       csvCache.set(id, { headers: parsedHeaders, data: parsedData });
     } catch (error) {
@@ -174,53 +173,3 @@ const csvCache = new Map<
   string,
   { headers: string[]; data: Record<string, string>[] }
 >();
-
-export function parseCSV(text: string): string[][] {
-  const rows: string[][] = [];
-  let field = "";
-  let fields: string[] = [];
-  let inQuotes = false;
-
-  for (let i = 0; i < text.length; i++) {
-    const char = text[i];
-
-    if (inQuotes) {
-      if (char === '"') {
-        if (i + 1 < text.length && text[i + 1] === '"') {
-          field += '"';
-          i++;
-        } else {
-          inQuotes = false;
-        }
-      } else {
-        field += char;
-      }
-    } else if (char === '"') {
-      inQuotes = true;
-    } else if (char === ",") {
-      fields.push(field);
-      field = "";
-    } else if (char === "\n" || char === "\r") {
-      if (char === "\r" && i + 1 < text.length && text[i + 1] === "\n") {
-        i++;
-      }
-      fields.push(field);
-      field = "";
-      rows.push(fields);
-      fields = [];
-    } else {
-      field += char;
-    }
-  }
-
-  if (inQuotes) {
-    throw new Error("Malformed CSV: unterminated quoted field");
-  }
-
-  if (field.length > 0 || fields.length > 0) {
-    fields.push(field);
-    rows.push(fields);
-  }
-
-  return rows;
-}

web/src/components/tools/parseCSV.test.ts

@@ -1,84 +0,0 @@
-import { parseCSV } from "./CSVContent";
-
-describe("parseCSV", () => {
-  it("parses simple comma-separated rows", () => {
-    expect(parseCSV("a,b,c\n1,2,3")).toEqual([
-      ["a", "b", "c"],
-      ["1", "2", "3"],
-    ]);
-  });
-
-  it("preserves commas inside quoted fields", () => {
-    expect(parseCSV('name,address\nAlice,"123 Main St, Apt 4"')).toEqual([
-      ["name", "address"],
-      ["Alice", "123 Main St, Apt 4"],
-    ]);
-  });
-
-  it("handles escaped double quotes inside quoted fields", () => {
-    expect(parseCSV('a,b\n"say ""hello""",world')).toEqual([
-      ["a", "b"],
-      ['say "hello"', "world"],
-    ]);
-  });
-
-  it("handles newlines inside quoted fields", () => {
-    expect(parseCSV('a,b\n"line1\nline2",val')).toEqual([
-      ["a", "b"],
-      ["line1\nline2", "val"],
-    ]);
-  });
-
-  it("handles CRLF line endings", () => {
-    expect(parseCSV("a,b\r\n1,2\r\n3,4")).toEqual([
-      ["a", "b"],
-      ["1", "2"],
-      ["3", "4"],
-    ]);
-  });
-
-  it("handles empty fields", () => {
-    expect(parseCSV("a,b,c\n1,,3")).toEqual([
-      ["a", "b", "c"],
-      ["1", "", "3"],
-    ]);
-  });
-
-  it("handles a single element", () => {
-    expect(parseCSV("a")).toEqual([["a"]]);
-  });
-
-  it("handles a single row with no newline", () => {
-    expect(parseCSV("a,b,c")).toEqual([["a", "b", "c"]]);
-  });
-
-  it("handles quoted fields that are entirely empty", () => {
-    expect(parseCSV('a,b\n"",val')).toEqual([
-      ["a", "b"],
-      ["", "val"],
-    ]);
-  });
-
-  it("handles multiple quoted fields with commas", () => {
-    expect(parseCSV('"foo, bar","baz, qux"\n"1, 2","3, 4"')).toEqual([
-      ["foo, bar", "baz, qux"],
-      ["1, 2", "3, 4"],
-    ]);
-  });
-
-  it("throws on unterminated quoted field", () => {
-    expect(() => parseCSV('a,b\n"foo,bar')).toThrow(
-      "Malformed CSV: unterminated quoted field"
-    );
-  });
-
-  it("throws on unterminated quote at end of input", () => {
-    expect(() => parseCSV('"unterminated')).toThrow(
-      "Malformed CSV: unterminated quoted field"
-    );
-  });
-
-  it("returns empty array for empty input", () => {
-    expect(parseCSV("")).toEqual([]);
-  });
-});

web/src/ee/providers/AppModeProvider.tsx

@@ -0,0 +1,55 @@
+"use client";
+
+import React, { useState, useCallback, useEffect } from "react";
+import { usePaidEnterpriseFeaturesEnabled } from "@/components/settings/usePaidEnterpriseFeaturesEnabled";
+import { AppModeContext, AppMode } from "@/providers/AppModeProvider";
+import { useUser } from "@/providers/UserProvider";
+import { useSettingsContext } from "@/providers/SettingsProvider";
+
+export interface AppModeProviderProps {
+  children: React.ReactNode;
+}
+
+/**
+ * Provider for application mode (Search/Chat).
+ *
+ * This controls how user queries are handled:
+ * - **search**: Forces search mode - quick document lookup
+ * - **chat**: Forces chat mode - conversation with follow-up questions
+ *
+ * The initial mode is read from the user's persisted `default_app_mode` preference.
+ * When search mode is unavailable (admin setting or no connectors), the mode is locked to "chat".
+ */
+export function AppModeProvider({ children }: AppModeProviderProps) {
+  const isPaidEnterpriseFeaturesEnabled = usePaidEnterpriseFeaturesEnabled();
+  const { user } = useUser();
+  const { isSearchModeAvailable } = useSettingsContext();
+
+  const persistedMode = user?.preferences?.default_app_mode;
+  const [appMode, setAppModeState] = useState<AppMode>("chat");
+
+  useEffect(() => {
+    if (!isPaidEnterpriseFeaturesEnabled || !isSearchModeAvailable) {
+      setAppModeState("chat");
+      return;
+    }
+
+    if (persistedMode) {
+      setAppModeState(persistedMode.toLowerCase() as AppMode);
+    }
+  }, [isPaidEnterpriseFeaturesEnabled, isSearchModeAvailable, persistedMode]);
+
+  const setAppMode = useCallback(
+    (mode: AppMode) => {
+      if (!isPaidEnterpriseFeaturesEnabled || !isSearchModeAvailable) return;
+      setAppModeState(mode);
+    },
+    [isPaidEnterpriseFeaturesEnabled, isSearchModeAvailable]
+  );
+
+  return (
+    <AppModeContext.Provider value={{ appMode, setAppMode }}>
+      {children}
+    </AppModeContext.Provider>
+  );
+}

web/src/ee/providers/QueryControllerProvider.tsx

@@ -8,15 +8,14 @@ import {
   SearchFullResponse,
 } from "@/lib/search/interfaces";
 import { classifyQuery, searchDocuments } from "@/ee/lib/search/svc";
+import { useAppMode } from "@/providers/AppModeProvider";
 import useAppFocus from "@/hooks/useAppFocus";
 import { usePaidEnterpriseFeaturesEnabled } from "@/components/settings/usePaidEnterpriseFeaturesEnabled";
 import { useSettingsContext } from "@/providers/SettingsProvider";
-import { useUser } from "@/providers/UserProvider";
 import {
   QueryControllerContext,
+  QueryClassification,
   QueryControllerValue,
-  QueryState,
-  AppMode,
 } from "@/providers/QueryControllerProvider";
 
 interface QueryControllerProviderProps {
@@ -26,53 +25,19 @@ interface QueryControllerProviderProps {
 export function QueryControllerProvider({
   children,
 }: QueryControllerProviderProps) {
+  const { appMode, setAppMode } = useAppMode();
   const appFocus = useAppFocus();
   const isPaidEnterpriseFeaturesEnabled = usePaidEnterpriseFeaturesEnabled();
   const settings = useSettingsContext();
   const { isSearchModeAvailable: searchUiEnabled } = settings;
-  const { user } = useUser();
 
-  // ── Merged query state (discriminated union) ──────────────────────────
-  const [state, setState] = useState<QueryState>({
-    phase: "idle",
-    appMode: "chat",
-  });
-
-  // Persistent app-mode preference — survives phase transitions and is
-  // used to restore the correct mode when resetting back to idle.
-  const appModeRef = useRef<AppMode>("chat");
-
-  // ── App mode sync from user preferences ───────────────────────────────
-  const persistedMode = user?.preferences?.default_app_mode;
-
-  useEffect(() => {
-    let mode: AppMode = "chat";
-    if (isPaidEnterpriseFeaturesEnabled && searchUiEnabled && persistedMode) {
-      const lower = persistedMode.toLowerCase();
-      mode = (["auto", "search", "chat"] as const).includes(lower as AppMode)
-        ? (lower as AppMode)
-        : "chat";
-    }
-    appModeRef.current = mode;
-    setState((prev) =>
-      prev.phase === "idle" ? { phase: "idle", appMode: mode } : prev
-    );
-  }, [isPaidEnterpriseFeaturesEnabled, searchUiEnabled, persistedMode]);
-
-  const setAppMode = useCallback(
-    (mode: AppMode) => {
-      if (!isPaidEnterpriseFeaturesEnabled || !searchUiEnabled) return;
-      setState((prev) => {
-        if (prev.phase !== "idle") return prev;
-        appModeRef.current = mode;
-        return { phase: "idle", appMode: mode };
-      });
-    },
-    [isPaidEnterpriseFeaturesEnabled, searchUiEnabled]
-  );
-
-  // ── Ancillary state ───────────────────────────────────────────────────
+  // Query state
   const [query, setQuery] = useState<string | null>(null);
+  const [classification, setClassification] =
+    useState<QueryClassification>(null);
+  const [isClassifying, setIsClassifying] = useState(false);
+
+  // Search state
   const [searchResults, setSearchResults] = useState<SearchDocWithContent[]>(
     []
   );
@@ -86,7 +51,7 @@ export function QueryControllerProvider({
   const searchAbortRef = useRef<AbortController | null>(null);
 
   /**
-   * Perform document search (pure data-fetching, no phase side effects)
+   * Perform document search
    */
   const performSearch = useCallback(
     async (searchQuery: string, filters?: BaseFilters): Promise<void> => {
@@ -120,15 +85,19 @@ export function QueryControllerProvider({
         setLlmSelectedDocIds(response.llm_selected_doc_ids ?? null);
       } catch (err) {
         if (err instanceof Error && err.name === "AbortError") {
-          throw err;
+          return;
         }
 
         setError("Document search failed. Please try again.");
         setSearchResults([]);
         setLlmSelectedDocIds(null);
+      } finally {
+        // After we've performed a search, we automatically switch to "search" mode.
+        // This is a "sticky" implementation; on purpose.
+        setAppMode("search");
       }
     },
-    []
+    [setAppMode]
   );
 
   /**
@@ -143,6 +112,8 @@ export function QueryControllerProvider({
       const controller = new AbortController();
       classifyAbortRef.current = controller;
 
+      setIsClassifying(true);
+
       try {
         const response: SearchFlowClassificationResponse = await classifyQuery(
           classifyQueryText,
@@ -158,6 +129,8 @@ export function QueryControllerProvider({
 
         setError("Query classification failed. Falling back to chat.");
         return "chat";
+      } finally {
+        setIsClassifying(false);
       }
     },
     []
@@ -175,51 +148,62 @@ export function QueryControllerProvider({
       setQuery(submitQuery);
       setError(null);
 
-      const currentAppMode = appModeRef.current;
-
-      // Always route through chat if:
-      // 1. Not Enterprise Enabled
-      // 2. Admin has disabled the Search UI
-      // 3. Not in the "New Session" tab
-      // 4. In "New Session" tab but app-mode is "Chat"
+      // 1.
+      // We always route through chat if we're not Enterprise Enabled.
+      //
+      // 2.
+      // We always route through chat if the admin has disabled the Search UI.
+      //
+      // 3.
+      // We only go down the classification route if we're in the "New Session" tab.
+      // Everywhere else, we always use the chat-flow.
+      //
+      // 4.
+      // If we're in the "New Session" tab and the app-mode is "Chat", we continue with the chat-flow anyways.
       if (
         !isPaidEnterpriseFeaturesEnabled ||
         !searchUiEnabled ||
         !appFocus.isNewSession() ||
-        currentAppMode === "chat"
+        appMode === "chat"
       ) {
-        setState({ phase: "chat" });
+        setClassification("chat");
         setSearchResults([]);
         setLlmSelectedDocIds(null);
         onChat(submitQuery);
         return;
       }
 
-      // Search mode: immediately show SearchUI with loading state
-      if (currentAppMode === "search") {
-        setState({ phase: "searching" });
-        try {
-          await performSearch(submitQuery, filters);
-        } catch (err) {
-          if (err instanceof Error && err.name === "AbortError") return;
-          throw err;
-        }
-        setState({ phase: "search-results" });
+      if (appMode === "search") {
+        await performSearch(submitQuery, filters);
+        setClassification("search");
         return;
       }
 
+      // # Note (@raunakab)
+      //
+      // Interestingly enough, for search, we do:
+      // 1. setClassification("search")
+      // 2. performSearch
+      //
+      // But for chat, we do:
+      // 1. performChat
+      // 2. setClassification("chat")
+      //
+      // The ChatUI has a nice loading UI, so it's fine for us to prematurely set the
+      // classification-state before the chat has finished loading.
+      //
+      // However, the SearchUI does not. Prematurely setting the classification-state
+      // will lead to a slightly ugly UI.
+
       // Auto mode: classify first, then route
-      setState({ phase: "classifying" });
       try {
         const result = await performClassification(submitQuery);
 
         if (result === "search") {
-          setState({ phase: "searching" });
           await performSearch(submitQuery, filters);
-          setState({ phase: "search-results" });
-          appModeRef.current = "search";
+          setClassification("search");
         } else {
-          setState({ phase: "chat" });
+          setClassification("chat");
           setSearchResults([]);
           setLlmSelectedDocIds(null);
           onChat(submitQuery);
@@ -229,13 +213,14 @@ export function QueryControllerProvider({
           return;
         }
 
-        setState({ phase: "chat" });
+        setClassification("chat");
         setSearchResults([]);
         setLlmSelectedDocIds(null);
         onChat(submitQuery);
       }
     },
     [
+      appMode,
       appFocus,
       performClassification,
       performSearch,
@@ -250,14 +235,7 @@ export function QueryControllerProvider({
   const refineSearch = useCallback(
     async (filters: BaseFilters): Promise<void> => {
       if (!query) return;
-      setState({ phase: "searching" });
-      try {
-        await performSearch(query, filters);
-      } catch (err) {
-        if (err instanceof Error && err.name === "AbortError") return;
-        throw err;
-      }
-      setState({ phase: "search-results" });
+      await performSearch(query, filters);
     },
     [query, performSearch]
   );
@@ -276,7 +254,7 @@ export function QueryControllerProvider({
     }
 
     setQuery(null);
-    setState({ phase: "idle", appMode: appModeRef.current });
+    setClassification(null);
     setSearchResults([]);
     setLlmSelectedDocIds(null);
     setError(null);
@@ -284,8 +262,8 @@ export function QueryControllerProvider({
 
   const value: QueryControllerValue = useMemo(
     () => ({
-      state,
-      setAppMode,
+      classification,
+      isClassifying,
       searchResults,
       llmSelectedDocIds,
       error,
@@ -294,8 +272,8 @@ export function QueryControllerProvider({
       reset,
     }),
     [
-      state,
-      setAppMode,
+      classification,
+      isClassifying,
       searchResults,
       llmSelectedDocIds,
       error,
@@ -305,7 +283,7 @@ export function QueryControllerProvider({
     ]
   );
 
-  // Sync state with navigation context
+  // Sync classification state with navigation context
   useEffect(reset, [appFocus, reset]);
 
   return (

web/src/ee/sections/SearchCard.tsx

@@ -56,7 +56,7 @@ export default function SearchCard({
 
   return (
     <Interactive.Stateless onClick={handleClick} prominence="secondary">
-      <Interactive.Container heightVariant="fit" widthVariant="full">
+      <Interactive.Container heightVariant="fit">
         <Section alignItems="start" gap={0} padding={0.25}>
           {/* Title Row */}
           <Section

web/src/ee/sections/SearchUI.tsx

@@ -18,17 +18,16 @@ import { getTimeFilterDate, TimeFilter } from "@/lib/time";
 import useTags from "@/hooks/useTags";
 import { SourceIcon } from "@/components/SourceIcon";
 import Text from "@/refresh-components/texts/Text";
+import LineItem from "@/refresh-components/buttons/LineItem";
 import { Section } from "@/layouts/general-layouts";
 import Popover, { PopoverMenu } from "@/refresh-components/Popover";
 import { SvgCheck, SvgClock, SvgTag } from "@opal/icons";
 import FilterButton from "@/refresh-components/buttons/FilterButton";
 import InputTypeIn from "@/refresh-components/inputs/InputTypeIn";
 import useFilter from "@/hooks/useFilter";
-import { LineItemButton } from "@opal/components";
 import { useQueryController } from "@/providers/QueryControllerProvider";
 import { cn } from "@/lib/utils";
 import { toast } from "@/hooks/useToast";
-import SimpleLoader from "@/refresh-components/loaders/SimpleLoader";
 
 // ============================================================================
 // Types
@@ -52,17 +51,22 @@ const TIME_FILTER_OPTIONS: { value: TimeFilter; label: string }[] = [
   { value: "year", label: "Past year" },
 ];
 
+// ============================================================================
+// SearchResults Component (default export)
+// ============================================================================
+
+/**
+ * Component for displaying search results with source filter sidebar.
+ */
 export default function SearchUI({ onDocumentClick }: SearchResultsProps) {
   // Available tags from backend
   const { tags: availableTags } = useTags();
   const {
-    state,
     searchResults: results,
     llmSelectedDocIds,
     error,
     refineSearch: onRefineSearch,
   } = useQueryController();
-
   const prevErrorRef = useRef<string | null>(null);
 
   // Show a toast notification when a new error occurs
@@ -193,15 +197,6 @@ export default function SearchUI({ onDocumentClick }: SearchResultsProps) {
 
   const showEmpty = !error && results.length === 0;
 
-  // Show a centered spinner while search is in-flight (after all hooks)
-  if (state.phase === "searching") {
-    return (
-      <div className="flex-1 min-h-0 w-full flex items-center justify-center">
-        <SimpleLoader />
-      </div>
-    );
-  }
-
   return (
     <div className="flex-1 min-h-0 w-full flex flex-col gap-3">
       {/* ── Top row: Filters + Result count ── */}
@@ -231,19 +226,18 @@ export default function SearchUI({ onDocumentClick }: SearchResultsProps) {
               <Popover.Content align="start" width="md">
                 <PopoverMenu>
                   {TIME_FILTER_OPTIONS.map((opt) => (
-                    <LineItemButton
+                    <LineItem
                       key={opt.value}
                       onClick={() => {
                         setTimeFilter(opt.value);
                         setTimeFilterOpen(false);
                         onRefineSearch(buildFilters({ time: opt.value }));
                       }}
-                      state={timeFilter === opt.value ? "selected" : "empty"}
+                      selected={timeFilter === opt.value}
                       icon={timeFilter === opt.value ? SvgCheck : SvgClock}
-                      title={opt.label}
-                      sizePreset="main-ui"
-                      variant="section"
-                    />
+                    >
+                      {opt.label}
+                    </LineItem>
                   ))}
                 </PopoverMenu>
               </Popover.Content>
@@ -284,7 +278,7 @@ export default function SearchUI({ onDocumentClick }: SearchResultsProps) {
                         t.tag_value === tag.tag_value
                     );
                     return (
-                      <LineItemButton
+                      <LineItem
                         key={`${tag.tag_key}=${tag.tag_value}`}
                         onClick={() => {
                           const next = isSelected
@@ -297,12 +291,11 @@ export default function SearchUI({ onDocumentClick }: SearchResultsProps) {
                           setSelectedTags(next);
                           onRefineSearch(buildFilters({ tags: next }));
                         }}
-                        state={isSelected ? "selected" : "empty"}
+                        selected={isSelected}
                         icon={isSelected ? SvgCheck : SvgTag}
-                        title={tag.tag_value}
-                        sizePreset="main-ui"
-                        variant="section"
-                      />
+                      >
+                        {tag.tag_value}
+                      </LineItem>
                     );
                   })}
                 </PopoverMenu>
@@ -364,7 +357,7 @@ export default function SearchUI({ onDocumentClick }: SearchResultsProps) {
           <div className="flex-1 min-h-0 overflow-y-auto flex flex-col gap-4 px-1">
             <Section gap={0.25} height="fit">
               {sourcesWithMeta.map(({ source, meta, count }) => (
-                <LineItemButton
+                <LineItem
                   key={source}
                   icon={(props) => (
                     <SourceIcon
@@ -374,15 +367,12 @@ export default function SearchUI({ onDocumentClick }: SearchResultsProps) {
                     />
                   )}
                   onClick={() => handleSourceToggle(source)}
-                  state={
-                    selectedSources.includes(source) ? "selected" : "empty"
-                  }
-                  title={meta.displayName}
-                  selectVariant="select-heavy"
-                  sizePreset="main-ui"
-                  variant="section"
+                  selected={selectedSources.includes(source)}
+                  emphasized
                   rightChildren={<Text text03>{count}</Text>}
-                />
+                >
+                  {meta.displayName}
+                </LineItem>
               ))}
             </Section>
           </div>

web/src/hooks/useAdminUsers.ts

@@ -1,123 +0,0 @@
-"use client";
-
-import { useCallback } from "react";
-import useSWR from "swr";
-import { errorHandlingFetcher } from "@/lib/fetcher";
-import { NEXT_PUBLIC_CLOUD_ENABLED } from "@/lib/constants";
-import { UserStatus } from "@/lib/types";
-import type { UserRole, InvitedUserSnapshot } from "@/lib/types";
-import type {
-  UserRow,
-  UserGroupInfo,
-} from "@/refresh-pages/admin/UsersPage/interfaces";
-
-// ---------------------------------------------------------------------------
-// Backend response shape (GET /manage/users/accepted/all)
-// ---------------------------------------------------------------------------
-
-interface FullUserSnapshot {
-  id: string;
-  email: string;
-  role: UserRole;
-  is_active: boolean;
-  password_configured: boolean;
-  personal_name: string | null;
-  created_at: string;
-  updated_at: string;
-  groups: UserGroupInfo[];
-  is_scim_synced: boolean;
-}
-
-// ---------------------------------------------------------------------------
-// Converters
-// ---------------------------------------------------------------------------
-
-function toUserRow(snapshot: FullUserSnapshot): UserRow {
-  return {
-    id: snapshot.id,
-    email: snapshot.email,
-    role: snapshot.role,
-    status: snapshot.is_active ? UserStatus.ACTIVE : UserStatus.INACTIVE,
-    is_active: snapshot.is_active,
-    is_scim_synced: snapshot.is_scim_synced,
-    personal_name: snapshot.personal_name,
-    created_at: snapshot.created_at,
-    updated_at: snapshot.updated_at,
-    groups: snapshot.groups,
-  };
-}
-
-function emailToUserRow(
-  email: string,
-  status: UserStatus.INVITED | UserStatus.REQUESTED
-): UserRow {
-  return {
-    id: null,
-    email,
-    role: null,
-    status,
-    is_active: false,
-    is_scim_synced: false,
-    personal_name: null,
-    created_at: null,
-    updated_at: null,
-    groups: [],
-  };
-}
-
-// ---------------------------------------------------------------------------
-// Hook
-// ---------------------------------------------------------------------------
-
-export default function useAdminUsers() {
-  const {
-    data: acceptedData,
-    isLoading: acceptedLoading,
-    error: acceptedError,
-    mutate: acceptedMutate,
-  } = useSWR<FullUserSnapshot[]>(
-    "/api/manage/users/accepted/all",
-    errorHandlingFetcher
-  );
-
-  const {
-    data: invitedData,
-    isLoading: invitedLoading,
-    error: invitedError,
-    mutate: invitedMutate,
-  } = useSWR<InvitedUserSnapshot[]>(
-    "/api/manage/users/invited",
-    errorHandlingFetcher
-  );
-
-  const {
-    data: requestedData,
-    isLoading: requestedLoading,
-    error: requestedError,
-    mutate: requestedMutate,
-  } = useSWR<InvitedUserSnapshot[]>(
-    NEXT_PUBLIC_CLOUD_ENABLED ? "/api/tenants/users/pending" : null,
-    errorHandlingFetcher
-  );
-
-  const acceptedRows = (acceptedData ?? []).map(toUserRow);
-  const invitedRows = (invitedData ?? []).map((u) =>
-    emailToUserRow(u.email, UserStatus.INVITED)
-  );
-  const requestedRows = (requestedData ?? []).map((u) =>
-    emailToUserRow(u.email, UserStatus.REQUESTED)
-  );
-
-  const users = [...invitedRows, ...requestedRows, ...acceptedRows];
-
-  const isLoading = acceptedLoading || invitedLoading || requestedLoading;
-  const error = acceptedError ?? invitedError ?? requestedError;
-
-  const refresh = useCallback(() => {
-    acceptedMutate();
-    invitedMutate();
-    requestedMutate();
-  }, [acceptedMutate, invitedMutate, requestedMutate]);
-
-  return { users, isLoading, error, refresh };
-}

web/src/hooks/useAppFocus.ts

@@ -5,7 +5,6 @@
 //
 // This is useful in determining what `SidebarTab` should be active, for example.
 
-import { useMemo } from "react";
 import { SEARCH_PARAM_NAMES } from "@/app/app/services/searchParams";
 import { usePathname, useSearchParams } from "next/navigation";
 
@@ -67,25 +66,31 @@ export default function useAppFocus(): AppFocus {
   const pathname = usePathname();
   const searchParams = useSearchParams();
 
-  const chatId = searchParams.get(SEARCH_PARAM_NAMES.CHAT_ID);
-  const agentId = searchParams.get(SEARCH_PARAM_NAMES.PERSONA_ID);
-  const projectId = searchParams.get(SEARCH_PARAM_NAMES.PROJECT_ID);
+  // Check if we're viewing a shared chat
+  if (pathname.startsWith("/app/shared/")) {
+    return new AppFocus("shared-chat");
+  }
 
-  // Memoize on the values that determine which AppFocus is constructed.
-  // AppFocus is immutable, so same inputs → same instance.
-  return useMemo(() => {
-    if (pathname.startsWith("/app/shared/")) {
-      return new AppFocus("shared-chat");
-    }
-    if (pathname.startsWith("/app/settings")) {
-      return new AppFocus("user-settings");
-    }
-    if (pathname.startsWith("/app/agents")) {
-      return new AppFocus("more-agents");
-    }
-    if (chatId) return new AppFocus({ type: "chat", id: chatId });
-    if (agentId) return new AppFocus({ type: "agent", id: agentId });
-    if (projectId) return new AppFocus({ type: "project", id: projectId });
-    return new AppFocus("new-session");
-  }, [pathname, chatId, agentId, projectId]);
+  // Check if we're on the user settings page
+  if (pathname.startsWith("/app/settings")) {
+    return new AppFocus("user-settings");
+  }
+
+  // Check if we're on the agents page
+  if (pathname.startsWith("/app/agents")) {
+    return new AppFocus("more-agents");
+  }
+
+  // Check search params for chat, agent, or project
+  const chatId = searchParams.get(SEARCH_PARAM_NAMES.CHAT_ID);
+  if (chatId) return new AppFocus({ type: "chat", id: chatId });
+
+  const agentId = searchParams.get(SEARCH_PARAM_NAMES.PERSONA_ID);
+  if (agentId) return new AppFocus({ type: "agent", id: agentId });
+
+  const projectId = searchParams.get(SEARCH_PARAM_NAMES.PROJECT_ID);
+  if (projectId) return new AppFocus({ type: "project", id: projectId });
+
+  // No search params means we're on a new session
+  return new AppFocus("new-session");
 }

web/src/hooks/useContainerCenter.ts

@@ -38,7 +38,7 @@ function measure(el: HTMLElement): { x: number; y: number } | null {
  */
 export default function useContainerCenter(): ContainerCenter {
   const pathname = usePathname();
-  const { isMediumScreen } = useScreenSize();
+  const { isSmallScreen } = useScreenSize();
   const [center, setCenter] = useState<{ x: number | null; y: number | null }>(
     () => {
       if (typeof document === "undefined") return NULL_CENTER;
@@ -68,9 +68,9 @@ export default function useContainerCenter(): ContainerCenter {
   }, [pathname]);
 
   return {
-    centerX: isMediumScreen ? null : center.x,
-    centerY: isMediumScreen ? null : center.y,
-    hasContainerCenter: isMediumScreen
+    centerX: isSmallScreen ? null : center.x,
+    centerY: isSmallScreen ? null : center.y,
+    hasContainerCenter: isSmallScreen
       ? false
       : center.x !== null && center.y !== null,
   };

web/src/hooks/useScreenSize.ts

@@ -2,7 +2,6 @@
 
 import {
   DESKTOP_SMALL_BREAKPOINT_PX,
-  DESKTOP_MEDIUM_BREAKPOINT_PX,
   MOBILE_SIDEBAR_BREAKPOINT_PX,
 } from "@/lib/constants";
 import { useState, useCallback } from "react";
@@ -13,7 +12,6 @@ export interface ScreenSize {
   width: number;
   isMobile: boolean;
   isSmallScreen: boolean;
-  isMediumScreen: boolean;
 }
 
 export default function useScreenSize(): ScreenSize {
@@ -36,13 +34,11 @@ export default function useScreenSize(): ScreenSize {
 
   const isMobile = sizes.width <= MOBILE_SIDEBAR_BREAKPOINT_PX;
   const isSmall = sizes.width <= DESKTOP_SMALL_BREAKPOINT_PX;
-  const isMedium = sizes.width <= DESKTOP_MEDIUM_BREAKPOINT_PX;
 
   return {
     height: sizes.height,
     width: sizes.width,
     isMobile: isMounted && isMobile,
     isSmallScreen: isMounted && isSmall,
-    isMediumScreen: isMounted && isMedium,
   };
 }

web/src/hooks/useUserCounts.ts

@@ -1,56 +0,0 @@
-"use client";
-
-import useSWR from "swr";
-import { errorHandlingFetcher } from "@/lib/fetcher";
-import type { InvitedUserSnapshot } from "@/lib/types";
-import { NEXT_PUBLIC_CLOUD_ENABLED } from "@/lib/constants";
-import type { StatusCountMap } from "@/refresh-pages/admin/UsersPage/interfaces";
-
-type UserCountsResponse = {
-  role_counts: Record<string, number>;
-  status_counts: Record<string, number>;
-};
-
-type UserCounts = {
-  activeCount: number | null;
-  invitedCount: number | null;
-  pendingCount: number | null;
-  roleCounts: Record<string, number>;
-  statusCounts: StatusCountMap;
-  refreshCounts: () => void;
-};
-
-export default function useUserCounts(): UserCounts {
-  const { data: countsData, mutate: refreshCounts } =
-    useSWR<UserCountsResponse>(
-      "/api/manage/users/counts",
-      errorHandlingFetcher
-    );
-
-  const { data: invitedUsers } = useSWR<InvitedUserSnapshot[]>(
-    "/api/manage/users/invited",
-    errorHandlingFetcher
-  );
-
-  const { data: pendingUsers } = useSWR<InvitedUserSnapshot[]>(
-    NEXT_PUBLIC_CLOUD_ENABLED ? "/api/tenants/users/pending" : null,
-    errorHandlingFetcher
-  );
-
-  const activeCount = countsData?.status_counts?.active ?? null;
-  const inactiveCount = countsData?.status_counts?.inactive ?? null;
-
-  return {
-    activeCount,
-    invitedCount: invitedUsers?.length ?? null,
-    pendingCount: pendingUsers?.length ?? null,
-    roleCounts: countsData?.role_counts ?? {},
-    statusCounts: {
-      ...(activeCount !== null ? { active: activeCount } : {}),
-      ...(inactiveCount !== null ? { inactive: inactiveCount } : {}),
-      ...(invitedUsers ? { invited: invitedUsers.length } : {}),
-      ...(pendingUsers ? { requested: pendingUsers.length } : {}),
-    } satisfies StatusCountMap,
-    refreshCounts,
-  };
-}

web/src/interfaces/llm.ts

@@ -7,7 +7,6 @@ export enum LLMProviderName {
   OPENROUTER = "openrouter",
   VERTEX_AI = "vertex_ai",
   BEDROCK = "bedrock",
-  LITELLM_PROXY = "litellm_proxy",
   CUSTOM = "custom",
 }
 
@@ -145,18 +144,6 @@ export interface OpenRouterFetchParams {
   provider_name?: string;
 }
 
-export interface LiteLLMProxyFetchParams {
-  api_base?: string;
-  api_key?: string;
-  provider_name?: string;
-  signal?: AbortSignal;
-}
-
-export interface LiteLLMProxyModelResponse {
-  provider_name: string;
-  model_name: string;
-}
-
 export interface VertexAIFetchParams {
   model_configurations?: ModelConfiguration[];
 }
@@ -166,5 +153,4 @@ export type FetchModelsParams =
   | OllamaFetchParams
   | LMStudioFetchParams
   | OpenRouterFetchParams
-  | LiteLLMProxyFetchParams
   | VertexAIFetchParams;

web/src/layouts/app-layouts.tsx

@@ -60,7 +60,7 @@ import {
 } from "@opal/icons";
 import MinimalMarkdown from "@/components/chat/MinimalMarkdown";
 import { useSettingsContext } from "@/providers/SettingsProvider";
-import type { AppMode } from "@/providers/QueryControllerProvider";
+import { AppMode, useAppMode } from "@/providers/AppModeProvider";
 import useAppFocus from "@/hooks/useAppFocus";
 import { useQueryController } from "@/providers/QueryControllerProvider";
 import { usePaidEnterpriseFeaturesEnabled } from "@/components/settings/usePaidEnterpriseFeaturesEnabled";
@@ -82,7 +82,7 @@ import useBrowserInfo from "@/hooks/useBrowserInfo";
  */
 function Header() {
   const isPaidEnterpriseFeaturesEnabled = usePaidEnterpriseFeaturesEnabled();
-  const { state, setAppMode } = useQueryController();
+  const { appMode, setAppMode } = useAppMode();
   const settings = useSettingsContext();
   const { isMobile } = useScreenSize();
   const { setFolded } = useAppSidebarContext();
@@ -108,6 +108,7 @@ function Header() {
     useChatSessions();
   const router = useRouter();
   const appFocus = useAppFocus();
+  const { classification } = useQueryController();
 
   const customHeaderContent =
     settings?.enterpriseSettings?.custom_header_content;
@@ -116,8 +117,7 @@ function Header() {
   // without this content still use.
   const pageWithHeaderContent = appFocus.isChat() || appFocus.isNewSession();
 
-  const effectiveMode: AppMode =
-    appFocus.isNewSession() && state.phase === "idle" ? state.appMode : "chat";
+  const effectiveMode: AppMode = appFocus.isNewSession() ? appMode : "chat";
 
   const availableProjects = useMemo(() => {
     if (!projects) return [];
@@ -323,7 +323,7 @@ function Header() {
           {isPaidEnterpriseFeaturesEnabled &&
             settings.isSearchModeAvailable &&
             appFocus.isNewSession() &&
-            state.phase === "idle" && (
+            !classification && (
               <Popover open={modePopoverOpen} onOpenChange={setModePopoverOpen}>
                 <Popover.Trigger asChild>
                   <OpenButton

web/src/layouts/settings-layouts.tsx

@@ -230,7 +230,7 @@ function SettingsHeader({
         </div>
       )}
 
-      <Spacer vertical rem={2.5} />
+      <Spacer vertical rem={1} />
 
       <div className="flex flex-col gap-6 px-4">
         <div className="flex w-full justify-between">

web/src/lib/admin-routes.ts

@@ -58,7 +58,6 @@ export const ADMIN_PATHS = {
   DOCUMENT_PROCESSING: "/admin/configuration/document-processing",
   KNOWLEDGE_GRAPH: "/admin/kg",
   USERS: "/admin/users",
-  USERS_V2: "/admin/users2",
   API_KEYS: "/admin/api-key",
   TOKEN_RATE_LIMITS: "/admin/token-rate-limits",
   USAGE: "/admin/performance/usage",
@@ -191,11 +190,6 @@ export const ADMIN_ROUTE_CONFIG: Record<string, AdminRouteConfig> = {
     title: "Manage Users",
     sidebarLabel: "Users",
   },
-  [ADMIN_PATHS.USERS_V2]: {
-    icon: SvgUser,
-    title: "Users & Requests",
-    sidebarLabel: "Users v2",
-  },
   [ADMIN_PATHS.API_KEYS]: {
     icon: SvgKey,
     title: "API Keys",

web/src/lib/constants.ts

@@ -123,7 +123,6 @@ export const MAX_FILES_TO_SHOW = 3;
 // SIZES
 export const MOBILE_SIDEBAR_BREAKPOINT_PX = 640;
 export const DESKTOP_SMALL_BREAKPOINT_PX = 912;
-export const DESKTOP_MEDIUM_BREAKPOINT_PX = 1232;
 export const DEFAULT_AGENT_AVATAR_SIZE_PX = 18;
 export const HORIZON_DISTANCE_PX = 800;
 export const LOGO_FOLDED_SIZE_PX = 24;