test: no vector db user file processing (#8854)

Co-authored-by: Raunak Bhagat <r@rabh.io>

.github/workflows/post-merge-beta-cherry-pick.yml

@@ -114,8 +114,10 @@ jobs:
 
       - name: Mark workflow as failed if cherry-pick failed
         if: steps.gate.outputs.should_cherrypick == 'true' && steps.run_cherry_pick.outputs.status == 'failure'
+        env:
+          CHERRY_PICK_REASON: ${{ steps.run_cherry_pick.outputs.reason }}
         run: |
-          echo "::error::Automated cherry-pick failed (${{ steps.run_cherry_pick.outputs.reason }})."
+          echo "::error::Automated cherry-pick failed (${CHERRY_PICK_REASON})."
           exit 1
 
   notify-slack-on-cherry-pick-failure:

.github/workflows/pr-playwright-tests.yml

@@ -603,7 +603,7 @@ jobs:
       pull-requests: write
     steps:
       - name: Download visual diff summaries
-        uses: actions/download-artifact@95815c38cf2ff2164869cbab79da8d1f422bc89e # ratchet:actions/download-artifact@v4
+        uses: actions/download-artifact@37930b1c2abaa49bbe596cd826c3c89aef350131
         with:
           pattern: screenshot-diff-summary-*
           path: summaries/

backend/ee/onyx/server/seeding.py

@@ -20,6 +20,7 @@ from ee.onyx.server.enterprise_settings.store import (
 from ee.onyx.server.enterprise_settings.store import upload_logo
 from onyx.context.search.enums import RecencyBiasSetting
 from onyx.db.engine.sql_engine import get_session_with_current_tenant
+from onyx.db.llm import fetch_existing_llm_provider
 from onyx.db.llm import update_default_provider
 from onyx.db.llm import upsert_llm_provider
 from onyx.db.models import Tool
@@ -117,15 +118,38 @@ def _seed_custom_tools(db_session: Session, tools: List[CustomToolSeed]) -> None
 def _seed_llms(
     db_session: Session, llm_upsert_requests: list[LLMProviderUpsertRequest]
 ) -> None:
-    if llm_upsert_requests:
-        logger.notice("Seeding LLMs")
-        seeded_providers = [
-            upsert_llm_provider(llm_upsert_request, db_session)
-            for llm_upsert_request in llm_upsert_requests
-        ]
-        update_default_provider(
-            provider_id=seeded_providers[0].id, db_session=db_session
-        )
+    if not llm_upsert_requests:
+        return
+
+    logger.notice("Seeding LLMs")
+    for request in llm_upsert_requests:
+        existing = fetch_existing_llm_provider(name=request.name, db_session=db_session)
+        if existing:
+            request.id = existing.id
+    seeded_providers = [
+        upsert_llm_provider(llm_upsert_request, db_session)
+        for llm_upsert_request in llm_upsert_requests
+    ]
+
+    default_provider = next(
+        (p for p in seeded_providers if p.model_configurations), None
+    )
+    if not default_provider:
+        return
+
+    visible_configs = [
+        mc for mc in default_provider.model_configurations if mc.is_visible
+    ]
+    default_config = (
+        visible_configs[0]
+        if visible_configs
+        else default_provider.model_configurations[0]
+    )
+    update_default_provider(
+        provider_id=default_provider.id,
+        model_name=default_config.name,
+        db_session=db_session,
+    )
 
 
 def _seed_personas(db_session: Session, personas: list[PersonaUpsertRequest]) -> None:

backend/ee/onyx/server/settings/api.py

@@ -109,6 +109,12 @@ def apply_license_status_to_settings(settings: Settings) -> Settings:
             if metadata.status == _BLOCKING_STATUS:
                 settings.application_status = metadata.status
                 settings.ee_features_enabled = False
+            elif metadata.used_seats > metadata.seats:
+                # License is valid but seat limit exceeded
+                settings.application_status = ApplicationStatus.SEAT_LIMIT_EXCEEDED
+                settings.seat_count = metadata.seats
+                settings.used_seats = metadata.used_seats
+                settings.ee_features_enabled = True
             else:
                 # Has a valid license (GRACE_PERIOD/PAYMENT_REMINDER still allow EE features)
                 settings.ee_features_enabled = True

backend/ee/onyx/server/tenants/provisioning.py

@@ -33,6 +33,7 @@ from onyx.configs.constants import MilestoneRecordType
 from onyx.db.engine.sql_engine import get_session_with_shared_schema
 from onyx.db.engine.sql_engine import get_session_with_tenant
 from onyx.db.image_generation import create_default_image_gen_config_from_api_key
+from onyx.db.llm import fetch_existing_llm_provider
 from onyx.db.llm import update_default_provider
 from onyx.db.llm import upsert_cloud_embedding_provider
 from onyx.db.llm import upsert_llm_provider
@@ -302,12 +303,17 @@ def configure_default_api_keys(db_session: Session) -> None:
 
     has_set_default_provider = False
 
-    def _upsert(request: LLMProviderUpsertRequest) -> None:
+    def _upsert(request: LLMProviderUpsertRequest, default_model: str) -> None:
         nonlocal has_set_default_provider
         try:
+            existing = fetch_existing_llm_provider(
+                name=request.name, db_session=db_session
+            )
+            if existing:
+                request.id = existing.id
             provider = upsert_llm_provider(request, db_session)
             if not has_set_default_provider:
-                update_default_provider(provider.id, db_session)
+                update_default_provider(provider.id, default_model, db_session)
                 has_set_default_provider = True
         except Exception as e:
             logger.error(f"Failed to configure {request.provider} provider: {e}")
@@ -325,14 +331,13 @@ def configure_default_api_keys(db_session: Session) -> None:
             name="OpenAI",
             provider=OPENAI_PROVIDER_NAME,
             api_key=OPENAI_DEFAULT_API_KEY,
-            default_model_name=default_model_name,
             model_configurations=_build_model_configuration_upsert_requests(
                 OPENAI_PROVIDER_NAME, recommendations
             ),
             api_key_changed=True,
             is_auto_mode=True,
         )
-        _upsert(openai_provider)
+        _upsert(openai_provider, default_model_name)
 
         # Create default image generation config using the OpenAI API key
         try:
@@ -361,14 +366,13 @@ def configure_default_api_keys(db_session: Session) -> None:
             name="Anthropic",
             provider=ANTHROPIC_PROVIDER_NAME,
             api_key=ANTHROPIC_DEFAULT_API_KEY,
-            default_model_name=default_model_name,
             model_configurations=_build_model_configuration_upsert_requests(
                 ANTHROPIC_PROVIDER_NAME, recommendations
             ),
             api_key_changed=True,
             is_auto_mode=True,
         )
-        _upsert(anthropic_provider)
+        _upsert(anthropic_provider, default_model_name)
     else:
         logger.info(
             "ANTHROPIC_DEFAULT_API_KEY not set, skipping Anthropic provider configuration"
@@ -393,14 +397,13 @@ def configure_default_api_keys(db_session: Session) -> None:
             name="Google Vertex AI",
             provider=VERTEXAI_PROVIDER_NAME,
             custom_config=custom_config,
-            default_model_name=default_model_name,
             model_configurations=_build_model_configuration_upsert_requests(
                 VERTEXAI_PROVIDER_NAME, recommendations
             ),
             api_key_changed=True,
             is_auto_mode=True,
         )
-        _upsert(vertexai_provider)
+        _upsert(vertexai_provider, default_model_name)
     else:
         logger.info(
             "VERTEXAI_DEFAULT_CREDENTIALS not set, skipping Vertex AI provider configuration"
@@ -432,12 +435,11 @@ def configure_default_api_keys(db_session: Session) -> None:
             name="OpenRouter",
             provider=OPENROUTER_PROVIDER_NAME,
             api_key=OPENROUTER_DEFAULT_API_KEY,
-            default_model_name=default_model_name,
             model_configurations=model_configurations,
             api_key_changed=True,
             is_auto_mode=True,
         )
-        _upsert(openrouter_provider)
+        _upsert(openrouter_provider, default_model_name)
     else:
         logger.info(
             "OPENROUTER_DEFAULT_API_KEY not set, skipping OpenRouter provider configuration"

backend/onyx/background/celery/tasks/beat_schedule.py

@@ -241,8 +241,7 @@ _VECTOR_DB_BEAT_TASK_NAMES: set[str] = {
     "check-for-index-attempt-cleanup",
     "check-for-doc-permissions-sync",
     "check-for-external-group-sync",
-    "check-for-documents-for-opensearch-migration",
-    "migrate-documents-from-vespa-to-opensearch",
+    "migrate-chunks-from-vespa-to-opensearch",
 }
 
 if DISABLE_VECTOR_DB:

backend/onyx/background/celery/tasks/user_file_processing/tasks.py

@@ -414,34 +414,31 @@ def _process_user_file_with_indexing(
         raise RuntimeError(f"Indexing pipeline failed for user file {user_file_id}")
 
 
-@shared_task(
-    name=OnyxCeleryTask.PROCESS_SINGLE_USER_FILE,
-    bind=True,
-    ignore_result=True,
-)
-def process_single_user_file(
-    self: Task, *, user_file_id: str, tenant_id: str  # noqa: ARG001
+def _process_user_file_impl(
+    *, user_file_id: str, tenant_id: str, redis_locking: bool
 ) -> None:
-    task_logger.info(f"process_single_user_file - Starting id={user_file_id}")
+    """Core implementation for processing a single user file.
+
+    When redis_locking=True, acquires a per-file Redis lock and clears the
+    queued-key guard (Celery path).  When redis_locking=False, skips all Redis
+    operations (BackgroundTask path).
+    """
+    task_logger.info(f"_process_user_file_impl - Starting id={user_file_id}")
     start = time.monotonic()
 
-    redis_client = get_redis_client(tenant_id=tenant_id)
-
-    # Clear the "queued" guard set by the beat generator so that the next beat
-    # cycle can re-enqueue this file if it is still in PROCESSING state after
-    # this task completes or fails.
-    redis_client.delete(_user_file_queued_key(user_file_id))
-
-    file_lock: RedisLock = redis_client.lock(
-        _user_file_lock_key(user_file_id),
-        timeout=CELERY_USER_FILE_PROCESSING_LOCK_TIMEOUT,
-    )
-
-    if not file_lock.acquire(blocking=False):
-        task_logger.info(
-            f"process_single_user_file - Lock held, skipping user_file_id={user_file_id}"
+    file_lock: RedisLock | None = None
+    if redis_locking:
+        redis_client = get_redis_client(tenant_id=tenant_id)
+        redis_client.delete(_user_file_queued_key(user_file_id))
+        file_lock = redis_client.lock(
+            _user_file_lock_key(user_file_id),
+            timeout=CELERY_USER_FILE_PROCESSING_LOCK_TIMEOUT,
         )
-        return None
+        if file_lock is not None and not file_lock.acquire(blocking=False):
+            task_logger.info(
+                f"_process_user_file_impl - Lock held, skipping user_file_id={user_file_id}"
+            )
+            return
 
     documents: list[Document] = []
     try:
@@ -449,15 +446,15 @@ def process_single_user_file(
             uf = db_session.get(UserFile, _as_uuid(user_file_id))
             if not uf:
                 task_logger.warning(
-                    f"process_single_user_file - UserFile not found id={user_file_id}"
+                    f"_process_user_file_impl - UserFile not found id={user_file_id}"
                 )
-                return None
+                return
 
             if uf.status != UserFileStatus.PROCESSING:
                 task_logger.info(
-                    f"process_single_user_file - Skipping id={user_file_id} status={uf.status}"
+                    f"_process_user_file_impl - Skipping id={user_file_id} status={uf.status}"
                 )
-                return None
+                return
 
             connector = LocalFileConnector(
                 file_locations=[uf.file_id],
@@ -471,7 +468,6 @@ def process_single_user_file(
                         [doc for doc in batch if not isinstance(doc, HierarchyNode)]
                     )
 
-                # update the document id to userfile id in the documents
                 for document in documents:
                     document.id = str(user_file_id)
                     document.source = DocumentSource.USER_FILE
@@ -493,9 +489,8 @@ def process_single_user_file(
 
             except Exception as e:
                 task_logger.exception(
-                    f"process_single_user_file - Error processing file id={user_file_id} - {e.__class__.__name__}"
+                    f"_process_user_file_impl - Error processing file id={user_file_id} - {e.__class__.__name__}"
                 )
-                # don't update the status if the user file is being deleted
                 current_user_file = db_session.get(UserFile, _as_uuid(user_file_id))
                 if (
                     current_user_file
@@ -504,33 +499,42 @@ def process_single_user_file(
                     uf.status = UserFileStatus.FAILED
                     db_session.add(uf)
                     db_session.commit()
-                return None
+                return
 
         elapsed = time.monotonic() - start
         task_logger.info(
-            f"process_single_user_file - Finished id={user_file_id} docs={len(documents)} elapsed={elapsed:.2f}s"
+            f"_process_user_file_impl - Finished id={user_file_id} docs={len(documents)} elapsed={elapsed:.2f}s"
         )
-        return None
     except Exception as e:
-        # Attempt to mark the file as failed
         with get_session_with_current_tenant() as db_session:
             uf = db_session.get(UserFile, _as_uuid(user_file_id))
             if uf:
-                # don't update the status if the user file is being deleted
                 if uf.status != UserFileStatus.DELETING:
                     uf.status = UserFileStatus.FAILED
                 db_session.add(uf)
                 db_session.commit()
 
         task_logger.exception(
-            f"process_single_user_file - Error processing file id={user_file_id} - {e.__class__.__name__}"
+            f"_process_user_file_impl - Error processing file id={user_file_id} - {e.__class__.__name__}"
         )
-        return None
     finally:
-        if file_lock.owned():
+        if file_lock is not None and file_lock.owned():
             file_lock.release()
 
 
+@shared_task(
+    name=OnyxCeleryTask.PROCESS_SINGLE_USER_FILE,
+    bind=True,
+    ignore_result=True,
+)
+def process_single_user_file(
+    self: Task, *, user_file_id: str, tenant_id: str  # noqa: ARG001
+) -> None:
+    _process_user_file_impl(
+        user_file_id=user_file_id, tenant_id=tenant_id, redis_locking=True
+    )
+
+
 @shared_task(
     name=OnyxCeleryTask.CHECK_FOR_USER_FILE_DELETE,
     soft_time_limit=300,
@@ -581,36 +585,38 @@ def check_for_user_file_delete(self: Task, *, tenant_id: str) -> None:
     return None
 
 
-@shared_task(
-    name=OnyxCeleryTask.DELETE_SINGLE_USER_FILE,
-    bind=True,
-    ignore_result=True,
-)
-def process_single_user_file_delete(
-    self: Task, *, user_file_id: str, tenant_id: str  # noqa: ARG001
+def _delete_user_file_impl(
+    *, user_file_id: str, tenant_id: str, redis_locking: bool
 ) -> None:
-    """Process a single user file delete."""
-    task_logger.info(f"process_single_user_file_delete - Starting id={user_file_id}")
-    redis_client = get_redis_client(tenant_id=tenant_id)
-    file_lock: RedisLock = redis_client.lock(
-        _user_file_delete_lock_key(user_file_id),
-        timeout=CELERY_GENERIC_BEAT_LOCK_TIMEOUT,
-    )
-    if not file_lock.acquire(blocking=False):
-        task_logger.info(
-            f"process_single_user_file_delete - Lock held, skipping user_file_id={user_file_id}"
+    """Core implementation for deleting a single user file.
+
+    When redis_locking=True, acquires a per-file Redis lock (Celery path).
+    When redis_locking=False, skips Redis operations (BackgroundTask path).
+    """
+    task_logger.info(f"_delete_user_file_impl - Starting id={user_file_id}")
+
+    file_lock: RedisLock | None = None
+    if redis_locking:
+        redis_client = get_redis_client(tenant_id=tenant_id)
+        file_lock = redis_client.lock(
+            _user_file_delete_lock_key(user_file_id),
+            timeout=CELERY_GENERIC_BEAT_LOCK_TIMEOUT,
         )
-        return None
+        if file_lock is not None and not file_lock.acquire(blocking=False):
+            task_logger.info(
+                f"_delete_user_file_impl - Lock held, skipping user_file_id={user_file_id}"
+            )
+            return
+
     try:
         with get_session_with_current_tenant() as db_session:
             user_file = db_session.get(UserFile, _as_uuid(user_file_id))
             if not user_file:
                 task_logger.info(
-                    f"process_single_user_file_delete - User file not found id={user_file_id}"
+                    f"_delete_user_file_impl - User file not found id={user_file_id}"
                 )
-                return None
+                return
 
-            # 1) Delete vector DB chunks (skip when disabled)
             if not DISABLE_VECTOR_DB:
                 if MANAGED_VESPA:
                     httpx_init_vespa_pool(
@@ -648,7 +654,6 @@ def process_single_user_file_delete(
                         chunk_count=chunk_count,
                     )
 
-            # 2) Delete the user-uploaded file content from filestore (blob + metadata)
             file_store = get_default_file_store()
             try:
                 file_store.delete_file(user_file.file_id)
@@ -656,26 +661,33 @@ def process_single_user_file_delete(
                     user_file_id_to_plaintext_file_name(user_file.id)
                 )
             except Exception as e:
-                # This block executed only if the file is not found in the filestore
                 task_logger.exception(
-                    f"process_single_user_file_delete - Error deleting file id={user_file.id} - {e.__class__.__name__}"
+                    f"_delete_user_file_impl - Error deleting file id={user_file.id} - {e.__class__.__name__}"
                 )
 
-            # 3) Finally, delete the UserFile row
             db_session.delete(user_file)
             db_session.commit()
-            task_logger.info(
-                f"process_single_user_file_delete - Completed id={user_file_id}"
-            )
+            task_logger.info(f"_delete_user_file_impl - Completed id={user_file_id}")
     except Exception as e:
         task_logger.exception(
-            f"process_single_user_file_delete - Error processing file id={user_file_id} - {e.__class__.__name__}"
+            f"_delete_user_file_impl - Error processing file id={user_file_id} - {e.__class__.__name__}"
         )
-        return None
     finally:
-        if file_lock.owned():
+        if file_lock is not None and file_lock.owned():
             file_lock.release()
-    return None
+
+
+@shared_task(
+    name=OnyxCeleryTask.DELETE_SINGLE_USER_FILE,
+    bind=True,
+    ignore_result=True,
+)
+def process_single_user_file_delete(
+    self: Task, *, user_file_id: str, tenant_id: str  # noqa: ARG001
+) -> None:
+    _delete_user_file_impl(
+        user_file_id=user_file_id, tenant_id=tenant_id, redis_locking=True
+    )
 
 
 @shared_task(
@@ -747,32 +759,30 @@ def check_for_user_file_project_sync(self: Task, *, tenant_id: str) -> None:
     return None
 
 
-@shared_task(
-    name=OnyxCeleryTask.PROCESS_SINGLE_USER_FILE_PROJECT_SYNC,
-    bind=True,
-    ignore_result=True,
-)
-def process_single_user_file_project_sync(
-    self: Task, *, user_file_id: str, tenant_id: str  # noqa: ARG001
+def _project_sync_user_file_impl(
+    *, user_file_id: str, tenant_id: str, redis_locking: bool
 ) -> None:
-    """Process a single user file project sync."""
-    task_logger.info(
-        f"process_single_user_file_project_sync - Starting id={user_file_id}"
-    )
+    """Core implementation for syncing a user file's project/persona metadata.
 
-    redis_client = get_redis_client(tenant_id=tenant_id)
-    redis_client.delete(_user_file_project_sync_queued_key(user_file_id))
+    When redis_locking=True, acquires a per-file Redis lock and clears the
+    queued-key guard (Celery path).  When redis_locking=False, skips Redis
+    operations (BackgroundTask path).
+    """
+    task_logger.info(f"_project_sync_user_file_impl - Starting id={user_file_id}")
 
-    file_lock: RedisLock = redis_client.lock(
-        user_file_project_sync_lock_key(user_file_id),
-        timeout=CELERY_USER_FILE_PROJECT_SYNC_LOCK_TIMEOUT,
-    )
-
-    if not file_lock.acquire(blocking=False):
-        task_logger.info(
-            f"process_single_user_file_project_sync - Lock held, skipping user_file_id={user_file_id}"
+    file_lock: RedisLock | None = None
+    if redis_locking:
+        redis_client = get_redis_client(tenant_id=tenant_id)
+        redis_client.delete(_user_file_project_sync_queued_key(user_file_id))
+        file_lock = redis_client.lock(
+            user_file_project_sync_lock_key(user_file_id),
+            timeout=CELERY_USER_FILE_PROJECT_SYNC_LOCK_TIMEOUT,
         )
-        return None
+        if file_lock is not None and not file_lock.acquire(blocking=False):
+            task_logger.info(
+                f"_project_sync_user_file_impl - Lock held, skipping user_file_id={user_file_id}"
+            )
+            return
 
     try:
         with get_session_with_current_tenant() as db_session:
@@ -783,11 +793,10 @@ def process_single_user_file_project_sync(
             ).scalar_one_or_none()
             if not user_file:
                 task_logger.info(
-                    f"process_single_user_file_project_sync - User file not found id={user_file_id}"
+                    f"_project_sync_user_file_impl - User file not found id={user_file_id}"
                 )
-                return None
+                return
 
-            # Sync project metadata to vector DB (skip when disabled)
             if not DISABLE_VECTOR_DB:
                 if MANAGED_VESPA:
                     httpx_init_vespa_pool(
@@ -822,7 +831,7 @@ def process_single_user_file_project_sync(
                     )
 
             task_logger.info(
-                f"process_single_user_file_project_sync - User file id={user_file_id}"
+                f"_project_sync_user_file_impl - User file id={user_file_id}"
             )
 
             user_file.needs_project_sync = False
@@ -835,11 +844,21 @@ def process_single_user_file_project_sync(
 
     except Exception as e:
         task_logger.exception(
-            f"process_single_user_file_project_sync - Error syncing project for file id={user_file_id} - {e.__class__.__name__}"
+            f"_project_sync_user_file_impl - Error syncing project for file id={user_file_id} - {e.__class__.__name__}"
         )
-        return None
     finally:
-        if file_lock.owned():
+        if file_lock is not None and file_lock.owned():
             file_lock.release()
 
-    return None
+
+@shared_task(
+    name=OnyxCeleryTask.PROCESS_SINGLE_USER_FILE_PROJECT_SYNC,
+    bind=True,
+    ignore_result=True,
+)
+def process_single_user_file_project_sync(
+    self: Task, *, user_file_id: str, tenant_id: str  # noqa: ARG001
+) -> None:
+    _project_sync_user_file_impl(
+        user_file_id=user_file_id, tenant_id=tenant_id, redis_locking=True
+    )

backend/onyx/db/image_generation.py

@@ -202,7 +202,6 @@ def create_default_image_gen_config_from_api_key(
             api_key=api_key,
             api_base=None,
             api_version=None,
-            default_model_name=model_name,
             deployment_name=None,
             is_public=True,
         )

backend/onyx/db/llm.py

@@ -213,11 +213,29 @@ def upsert_llm_provider(
     llm_provider_upsert_request: LLMProviderUpsertRequest,
     db_session: Session,
 ) -> LLMProviderView:
-    existing_llm_provider = fetch_existing_llm_provider(
-        name=llm_provider_upsert_request.name, db_session=db_session
-    )
+    existing_llm_provider: LLMProviderModel | None = None
+    if llm_provider_upsert_request.id:
+        existing_llm_provider = fetch_existing_llm_provider_by_id(
+            id=llm_provider_upsert_request.id, db_session=db_session
+        )
+        if not existing_llm_provider:
+            raise ValueError(
+                f"LLM provider with id {llm_provider_upsert_request.id} not found"
+            )
 
-    if not existing_llm_provider:
+        if existing_llm_provider.name != llm_provider_upsert_request.name:
+            raise ValueError(
+                f"LLM provider with id {llm_provider_upsert_request.id} name change not allowed"
+            )
+    else:
+        existing_llm_provider = fetch_existing_llm_provider(
+            name=llm_provider_upsert_request.name, db_session=db_session
+        )
+        if existing_llm_provider:
+            raise ValueError(
+                f"LLM provider with name '{llm_provider_upsert_request.name}'"
+                " already exists"
+            )
         existing_llm_provider = LLMProviderModel(name=llm_provider_upsert_request.name)
         db_session.add(existing_llm_provider)
 
@@ -238,11 +256,7 @@ def upsert_llm_provider(
     existing_llm_provider.api_base = api_base
     existing_llm_provider.api_version = llm_provider_upsert_request.api_version
     existing_llm_provider.custom_config = custom_config
-    # TODO: Remove default model name on api change
-    # Needed due to /provider/{id}/default endpoint not disclosing the default model name
-    existing_llm_provider.default_model_name = (
-        llm_provider_upsert_request.default_model_name
-    )
+
     existing_llm_provider.is_public = llm_provider_upsert_request.is_public
     existing_llm_provider.is_auto_mode = llm_provider_upsert_request.is_auto_mode
     existing_llm_provider.deployment_name = llm_provider_upsert_request.deployment_name
@@ -306,15 +320,6 @@ def upsert_llm_provider(
                 display_name=model_config.display_name,
             )
 
-    default_model = fetch_default_model(db_session, LLMModelFlowType.CHAT)
-    if default_model and default_model.llm_provider_id == existing_llm_provider.id:
-        _update_default_model(
-            db_session=db_session,
-            provider_id=existing_llm_provider.id,
-            model=existing_llm_provider.default_model_name,
-            flow_type=LLMModelFlowType.CHAT,
-        )
-
     # Make sure the relationship table stays up to date
     update_group_llm_provider_relationships__no_commit(
         llm_provider_id=existing_llm_provider.id,
@@ -488,6 +493,22 @@ def fetch_existing_llm_provider(
     return provider_model
 
 
+def fetch_existing_llm_provider_by_id(
+    id: int, db_session: Session
+) -> LLMProviderModel | None:
+    provider_model = db_session.scalar(
+        select(LLMProviderModel)
+        .where(LLMProviderModel.id == id)
+        .options(
+            selectinload(LLMProviderModel.model_configurations),
+            selectinload(LLMProviderModel.groups),
+            selectinload(LLMProviderModel.personas),
+        )
+    )
+
+    return provider_model
+
+
 def fetch_embedding_provider(
     db_session: Session, provider_type: EmbeddingProvider
 ) -> CloudEmbeddingProviderModel | None:
@@ -604,22 +625,13 @@ def remove_llm_provider__no_commit(db_session: Session, provider_id: int) -> Non
     db_session.flush()
 
 
-def update_default_provider(provider_id: int, db_session: Session) -> None:
-    # Attempt to get the default_model_name from the provider first
-    # TODO: Remove default_model_name check
-    provider = db_session.scalar(
-        select(LLMProviderModel).where(
-            LLMProviderModel.id == provider_id,
-        )
-    )
-
-    if provider is None:
-        raise ValueError(f"LLM Provider with id={provider_id} does not exist")
-
+def update_default_provider(
+    provider_id: int, model_name: str, db_session: Session
+) -> None:
     _update_default_model(
         db_session,
         provider_id,
-        provider.default_model_name,  # type: ignore[arg-type]
+        model_name,
         LLMModelFlowType.CHAT,
     )
 
@@ -805,12 +817,6 @@ def sync_auto_mode_models(
             )
             changes += 1
 
-    # In Auto mode, default model is always set from GitHub config
-    default_model = llm_recommendations.get_default_model(provider.provider)
-    if default_model and provider.default_model_name != default_model.name:
-        provider.default_model_name = default_model.name
-        changes += 1
-
     db_session.commit()
     return changes
 

backend/onyx/main.py

@@ -37,6 +37,7 @@ from onyx.configs.app_configs import APP_HOST
 from onyx.configs.app_configs import APP_PORT
 from onyx.configs.app_configs import AUTH_RATE_LIMITING_ENABLED
 from onyx.configs.app_configs import AUTH_TYPE
+from onyx.configs.app_configs import DISABLE_VECTOR_DB
 from onyx.configs.app_configs import LOG_ENDPOINT_LATENCY
 from onyx.configs.app_configs import OAUTH_CLIENT_ID
 from onyx.configs.app_configs import OAUTH_CLIENT_SECRET
@@ -254,8 +255,38 @@ def include_auth_router_with_prefix(
     )
 
 
+def validate_no_vector_db_settings() -> None:
+    """Validate that DISABLE_VECTOR_DB is not combined with incompatible settings.
+
+    Raises RuntimeError if DISABLE_VECTOR_DB is set alongside MULTI_TENANT or ENABLE_CRAFT,
+    since these modes require infrastructure that is removed in no-vector-DB deployments.
+    """
+    if not DISABLE_VECTOR_DB:
+        return
+
+    if MULTI_TENANT:
+        raise RuntimeError(
+            "DISABLE_VECTOR_DB cannot be used with MULTI_TENANT. "
+            "Multi-tenant deployments require the vector database for "
+            "per-tenant document indexing and search. Run in single-tenant "
+            "mode when disabling the vector database."
+        )
+
+    from onyx.server.features.build.configs import ENABLE_CRAFT
+
+    if ENABLE_CRAFT:
+        raise RuntimeError(
+            "DISABLE_VECTOR_DB cannot be used with ENABLE_CRAFT. "
+            "Onyx Craft requires background workers for sandbox lifecycle "
+            "management, which are removed in no-vector-DB deployments. "
+            "Disable Craft (ENABLE_CRAFT=false) when disabling the vector database."
+        )
+
+
 @asynccontextmanager
 async def lifespan(app: FastAPI) -> AsyncGenerator[None, None]:  # noqa: ARG001
+    validate_no_vector_db_settings()
+
     # Set recursion limit
     if SYSTEM_RECURSION_LIMIT is not None:
         sys.setrecursionlimit(SYSTEM_RECURSION_LIMIT)

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

@@ -97,7 +97,6 @@ def _build_llm_provider_request(
             ),  # Only this from source
             api_base=api_base,  # From request
             api_version=api_version,  # From request
-            default_model_name=model_name,
             deployment_name=deployment_name,  # From request
             is_public=True,
             groups=[],
@@ -136,7 +135,6 @@ def _build_llm_provider_request(
         api_key=api_key,
         api_base=api_base,
         api_version=api_version,
-        default_model_name=model_name,
         deployment_name=deployment_name,
         is_public=True,
         groups=[],
@@ -168,7 +166,6 @@ def _create_image_gen_llm_provider__no_commit(
         api_key=provider_request.api_key,
         api_base=provider_request.api_base,
         api_version=provider_request.api_version,
-        default_model_name=provider_request.default_model_name,
         deployment_name=provider_request.deployment_name,
         is_public=provider_request.is_public,
         custom_config=provider_request.custom_config,

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

@@ -22,7 +22,10 @@ from onyx.auth.users import current_chat_accessible_user
 from onyx.db.engine.sql_engine import get_session
 from onyx.db.enums import LLMModelFlowType
 from onyx.db.llm import can_user_access_llm_provider
+from onyx.db.llm import fetch_default_llm_model
+from onyx.db.llm import fetch_default_vision_model
 from onyx.db.llm import fetch_existing_llm_provider
+from onyx.db.llm import fetch_existing_llm_provider_by_id
 from onyx.db.llm import fetch_existing_llm_providers
 from onyx.db.llm import fetch_existing_models
 from onyx.db.llm import fetch_persona_with_groups
@@ -52,11 +55,12 @@ 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 LLMCost
 from onyx.server.manage.llm.models import LLMProviderDescriptor
+from onyx.server.manage.llm.models import LLMProviderResponse
 from onyx.server.manage.llm.models import LLMProviderUpsertRequest
 from onyx.server.manage.llm.models import LLMProviderView
-from onyx.server.manage.llm.models import ModelConfigurationUpsertRequest
 from onyx.server.manage.llm.models import OllamaFinalModelResponse
 from onyx.server.manage.llm.models import OllamaModelDetails
 from onyx.server.manage.llm.models import OllamaModelsRequest
@@ -233,12 +237,9 @@ def test_llm_configuration(
 
     test_api_key = test_llm_request.api_key
     test_custom_config = test_llm_request.custom_config
-    if test_llm_request.name:
-        # NOTE: we are querying by name. we probably should be querying by an invariant id, but
-        # as it turns out the name is not editable in the UI and other code also keys off name,
-        # so we won't rock the boat just yet.
-        existing_provider = fetch_existing_llm_provider(
-            name=test_llm_request.name, db_session=db_session
+    if test_llm_request.id:
+        existing_provider = fetch_existing_llm_provider_by_id(
+            id=test_llm_request.id, db_session=db_session
         )
         if existing_provider:
             test_custom_config = _restore_masked_custom_config_values(
@@ -268,7 +269,7 @@ def test_llm_configuration(
 
     llm = get_llm(
         provider=test_llm_request.provider,
-        model=test_llm_request.default_model_name,
+        model=test_llm_request.model,
         api_key=test_api_key,
         api_base=test_llm_request.api_base,
         api_version=test_llm_request.api_version,
@@ -303,7 +304,7 @@ def list_llm_providers(
     include_image_gen: bool = Query(False),
     _: User = Depends(current_admin_user),
     db_session: Session = Depends(get_session),
-) -> list[LLMProviderView]:
+) -> LLMProviderResponse[LLMProviderView]:
     start_time = datetime.now(timezone.utc)
     logger.debug("Starting to fetch LLM providers")
 
@@ -328,7 +329,15 @@ def list_llm_providers(
     duration = (end_time - start_time).total_seconds()
     logger.debug(f"Completed fetching LLM providers in {duration:.2f} seconds")
 
-    return llm_provider_list
+    return LLMProviderResponse[LLMProviderView].from_models(
+        providers=llm_provider_list,
+        default_text=DefaultModel.from_model_config(
+            fetch_default_llm_model(db_session)
+        ),
+        default_vision=DefaultModel.from_model_config(
+            fetch_default_vision_model(db_session)
+        ),
+    )
 
 
 @admin_router.put("/provider")
@@ -344,18 +353,44 @@ def put_llm_provider(
     # validate request (e.g. if we're intending to create but the name already exists we should throw an error)
     # NOTE: may involve duplicate fetching to Postgres, but we're assuming SQLAlchemy is smart enough to cache
     # the result
-    existing_provider = fetch_existing_llm_provider(
+    existing_provider = None
+    if llm_provider_upsert_request.id:
+        existing_provider = fetch_existing_llm_provider_by_id(
+            id=llm_provider_upsert_request.id, db_session=db_session
+        )
+
+    # Check name constraints
+    # TODO: Once port from name to id is complete, unique name will no longer be required
+    if existing_provider and llm_provider_upsert_request.name != existing_provider.name:
+        raise HTTPException(
+            status_code=400,
+            detail="Renaming providers is not currently supported",
+        )
+
+    found_provider = fetch_existing_llm_provider(
         name=llm_provider_upsert_request.name, db_session=db_session
     )
+    if found_provider is not None and found_provider is not existing_provider:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Provider with name={llm_provider_upsert_request.name} already exists",
+        )
+
     if existing_provider and is_creation:
         raise HTTPException(
             status_code=400,
-            detail=f"LLM Provider with name {llm_provider_upsert_request.name} already exists",
+            detail=(
+                f"LLM Provider with name {llm_provider_upsert_request.name} and "
+                f"id={llm_provider_upsert_request.id} already exists"
+            ),
         )
     elif not existing_provider and not is_creation:
         raise HTTPException(
             status_code=400,
-            detail=f"LLM Provider with name {llm_provider_upsert_request.name} does not exist",
+            detail=(
+                f"LLM Provider with name {llm_provider_upsert_request.name} and "
+                f"id={llm_provider_upsert_request.id} does not exist"
+            ),
         )
 
     # SSRF Protection: Validate api_base and custom_config match stored values
@@ -393,22 +428,6 @@ def put_llm_provider(
                 deduplicated_personas.append(persona_id)
         llm_provider_upsert_request.personas = deduplicated_personas
 
-    default_model_found = False
-
-    for model_configuration in llm_provider_upsert_request.model_configurations:
-        if model_configuration.name == llm_provider_upsert_request.default_model_name:
-            model_configuration.is_visible = True
-            default_model_found = True
-
-    # TODO: Remove this logic on api change
-    # Believed to be a dead pathway but we want to be safe for now
-    if not default_model_found:
-        llm_provider_upsert_request.model_configurations.append(
-            ModelConfigurationUpsertRequest(
-                name=llm_provider_upsert_request.default_model_name, is_visible=True
-            )
-        )
-
     # the llm api key is sanitized when returned to clients, so the only time we
     # should get a real key is when it is explicitly changed
     if existing_provider and not llm_provider_upsert_request.api_key_changed:
@@ -438,8 +457,8 @@ def put_llm_provider(
             config = fetch_llm_recommendations_from_github()
             if config and llm_provider_upsert_request.provider in config.providers:
                 # Refetch the provider to get the updated model
-                updated_provider = fetch_existing_llm_provider(
-                    name=llm_provider_upsert_request.name, db_session=db_session
+                updated_provider = fetch_existing_llm_provider_by_id(
+                    id=result.id, db_session=db_session
                 )
                 if updated_provider:
                     sync_auto_mode_models(
@@ -469,28 +488,29 @@ def delete_llm_provider(
         raise HTTPException(status_code=404, detail=str(e))
 
 
-@admin_router.post("/provider/{provider_id}/default")
+@admin_router.post("/default")
 def set_provider_as_default(
-    provider_id: int,
+    default_model_request: DefaultModel,
     _: User = Depends(current_admin_user),
     db_session: Session = Depends(get_session),
 ) -> None:
-    update_default_provider(provider_id=provider_id, db_session=db_session)
+    update_default_provider(
+        provider_id=default_model_request.provider_id,
+        model_name=default_model_request.model_name,
+        db_session=db_session,
+    )
 
 
-@admin_router.post("/provider/{provider_id}/default-vision")
+@admin_router.post("/default-vision")
 def set_provider_as_default_vision(
-    provider_id: int,
-    vision_model: str | None = Query(
-        None, description="The default vision model to use"
-    ),
+    default_model: DefaultModel,
     _: User = Depends(current_admin_user),
     db_session: Session = Depends(get_session),
 ) -> None:
-    if vision_model is None:
-        raise HTTPException(status_code=404, detail="Vision model not provided")
     update_default_vision_provider(
-        provider_id=provider_id, vision_model=vision_model, db_session=db_session
+        provider_id=default_model.provider_id,
+        vision_model=default_model.model_name,
+        db_session=db_session,
     )
 
 
@@ -516,7 +536,7 @@ def get_auto_config(
 def get_vision_capable_providers(
     _: User = Depends(current_admin_user),
     db_session: Session = Depends(get_session),
-) -> list[VisionProviderResponse]:
+) -> LLMProviderResponse[VisionProviderResponse]:
     """Return a list of LLM providers and their models that support image input"""
     vision_models = fetch_existing_models(
         db_session=db_session, flow_types=[LLMModelFlowType.VISION]
@@ -545,7 +565,13 @@ def get_vision_capable_providers(
     ]
 
     logger.debug(f"Found {len(vision_provider_response)} vision-capable providers")
-    return vision_provider_response
+
+    return LLMProviderResponse[VisionProviderResponse].from_models(
+        providers=vision_provider_response,
+        default_vision=DefaultModel.from_model_config(
+            fetch_default_vision_model(db_session)
+        ),
+    )
 
 
 """Endpoints for all"""
@@ -555,7 +581,7 @@ def get_vision_capable_providers(
 def list_llm_provider_basics(
     user: User = Depends(current_chat_accessible_user),
     db_session: Session = Depends(get_session),
-) -> list[LLMProviderDescriptor]:
+) -> LLMProviderResponse[LLMProviderDescriptor]:
     """Get LLM providers accessible to the current user.
 
     Returns:
@@ -592,7 +618,15 @@ def list_llm_provider_basics(
         f"Completed fetching {len(accessible_providers)} user-accessible providers in {duration:.2f} seconds"
     )
 
-    return accessible_providers
+    return LLMProviderResponse[LLMProviderDescriptor].from_models(
+        providers=accessible_providers,
+        default_text=DefaultModel.from_model_config(
+            fetch_default_llm_model(db_session)
+        ),
+        default_vision=DefaultModel.from_model_config(
+            fetch_default_vision_model(db_session)
+        ),
+    )
 
 
 def get_valid_model_names_for_persona(
@@ -635,7 +669,7 @@ def list_llm_providers_for_persona(
     persona_id: int,
     user: User = Depends(current_chat_accessible_user),
     db_session: Session = Depends(get_session),
-) -> list[LLMProviderDescriptor]:
+) -> LLMProviderResponse[LLMProviderDescriptor]:
     """Get LLM providers for a specific persona.
 
     Returns providers that the user can access when using this persona:
@@ -682,7 +716,51 @@ def list_llm_providers_for_persona(
         f"Completed fetching {len(llm_provider_list)} LLM providers for persona {persona_id} in {duration:.2f} seconds"
     )
 
-    return llm_provider_list
+    # Get the default model and vision model for the persona
+    # TODO: Port persona's over to use ID
+    persona_default_provider = persona.llm_model_provider_override
+    persona_default_model = persona.llm_model_version_override
+
+    default_text_model = fetch_default_llm_model(db_session)
+    default_vision_model = fetch_default_vision_model(db_session)
+
+    # Build default_text and default_vision using persona overrides when available,
+    # falling back to the global defaults.
+    default_text = DefaultModel.from_model_config(default_text_model)
+    default_vision = DefaultModel.from_model_config(default_vision_model)
+
+    if persona_default_provider:
+        provider = fetch_existing_llm_provider(persona_default_provider, db_session)
+        if provider and can_user_access_llm_provider(
+            provider, user_group_ids, persona, is_admin=is_admin
+        ):
+            if persona_default_model:
+                # Persona specifies both provider and model — use them directly
+                default_text = DefaultModel(
+                    provider_id=provider.id,
+                    model_name=persona_default_model,
+                )
+            else:
+                # Persona specifies only the provider — pick a visible (public) model,
+                # falling back to any model on this provider
+                visible_model = next(
+                    (mc for mc in provider.model_configurations if mc.is_visible),
+                    None,
+                )
+                fallback_model = visible_model or next(
+                    iter(provider.model_configurations), None
+                )
+                if fallback_model:
+                    default_text = DefaultModel(
+                        provider_id=provider.id,
+                        model_name=fallback_model.name,
+                    )
+
+    return LLMProviderResponse[LLMProviderDescriptor].from_models(
+        providers=llm_provider_list,
+        default_text=default_text,
+        default_vision=default_vision,
+    )
 
 
 @admin_router.get("/provider-contextual-cost")

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

@@ -1,5 +1,9 @@
+from __future__ import annotations
+
 from typing import Any
+from typing import Generic
 from typing import TYPE_CHECKING
+from typing import TypeVar
 
 from pydantic import BaseModel
 from pydantic import Field
@@ -21,50 +25,22 @@ if TYPE_CHECKING:
         ModelConfiguration as ModelConfigurationModel,
     )
 
-
-# TODO: Clear this up on api refactor
-# There is still logic that requires sending each providers default model name
-# There is no logic that requires sending the providers default vision model name
-# We only send for the one that is actually the default
-def get_default_llm_model_name(llm_provider_model: "LLMProviderModel") -> str:
-    """Find the default conversation model name for a provider.
-
-    Returns the model name if found, otherwise returns empty string.
-    """
-    for model_config in llm_provider_model.model_configurations:
-        for flow in model_config.llm_model_flows:
-            if flow.is_default and flow.llm_model_flow_type == LLMModelFlowType.CHAT:
-                return model_config.name
-    return ""
-
-
-def get_default_vision_model_name(llm_provider_model: "LLMProviderModel") -> str | None:
-    """Find the default vision model name for a provider.
-
-    Returns the model name if found, otherwise returns None.
-    """
-    for model_config in llm_provider_model.model_configurations:
-        for flow in model_config.llm_model_flows:
-            if flow.is_default and flow.llm_model_flow_type == LLMModelFlowType.VISION:
-                return model_config.name
-    return None
+T = TypeVar("T", "LLMProviderDescriptor", "LLMProviderView", "VisionProviderResponse")
 
 
 class TestLLMRequest(BaseModel):
     # provider level
-    name: str | None = None
+    id: int | None = None
     provider: str
+    model: str
     api_key: str | None = None
     api_base: str | None = None
     api_version: str | None = None
     custom_config: dict[str, str] | None = None
 
     # model level
-    default_model_name: str
     deployment_name: str | None = None
 
-    model_configurations: list["ModelConfigurationUpsertRequest"]
-
     # if try and use the existing API/custom config key
     api_key_changed: bool
     custom_config_changed: bool
@@ -80,13 +56,10 @@ class LLMProviderDescriptor(BaseModel):
     """A descriptor for an LLM provider that can be safely viewed by
     non-admin users. Used when giving a list of available LLMs."""
 
+    id: int
     name: str
     provider: str
     provider_display_name: str  # Human-friendly name like "Claude (Anthropic)"
-    default_model_name: str
-    is_default_provider: bool | None
-    is_default_vision_provider: bool | None
-    default_vision_model: str | None
     model_configurations: list["ModelConfigurationView"]
 
     @classmethod
@@ -99,24 +72,12 @@ class LLMProviderDescriptor(BaseModel):
         )
 
         provider = llm_provider_model.provider
-        default_model_name = get_default_llm_model_name(llm_provider_model)
-        default_vision_model = get_default_vision_model_name(llm_provider_model)
-
-        is_default_provider = bool(default_model_name)
-        is_default_vision_provider = default_vision_model is not None
-
-        default_model_name = (
-            default_model_name or llm_provider_model.default_model_name or ""
-        )
 
         return cls(
+            id=llm_provider_model.id,
             name=llm_provider_model.name,
             provider=provider,
             provider_display_name=get_provider_display_name(provider),
-            default_model_name=default_model_name,
-            is_default_provider=is_default_provider,
-            is_default_vision_provider=is_default_vision_provider,
-            default_vision_model=default_vision_model,
             model_configurations=filter_model_configurations(
                 llm_provider_model.model_configurations, provider
             ),
@@ -130,18 +91,17 @@ class LLMProvider(BaseModel):
     api_base: str | None = None
     api_version: str | None = None
     custom_config: dict[str, str] | None = None
-    default_model_name: str
     is_public: bool = True
     is_auto_mode: bool = False
     groups: list[int] = Field(default_factory=list)
     personas: list[int] = Field(default_factory=list)
     deployment_name: str | None = None
-    default_vision_model: str | None = None
 
 
 class LLMProviderUpsertRequest(LLMProvider):
     # should only be used for a "custom" provider
     # for default providers, the built-in model names are used
+    id: int | None = None
     api_key_changed: bool = False
     custom_config_changed: bool = False
     model_configurations: list["ModelConfigurationUpsertRequest"] = []
@@ -157,8 +117,6 @@ class LLMProviderView(LLMProvider):
     """Stripped down representation of LLMProvider for display / limited access info only"""
 
     id: int
-    is_default_provider: bool | None = None
-    is_default_vision_provider: bool | None = None
     model_configurations: list["ModelConfigurationView"]
 
     @classmethod
@@ -180,16 +138,6 @@ class LLMProviderView(LLMProvider):
 
         provider = llm_provider_model.provider
 
-        default_model_name = get_default_llm_model_name(llm_provider_model)
-        default_vision_model = get_default_vision_model_name(llm_provider_model)
-
-        is_default_provider = bool(default_model_name)
-        is_default_vision_provider = default_vision_model is not None
-
-        default_model_name = (
-            default_model_name or llm_provider_model.default_model_name or ""
-        )
-
         return cls(
             id=llm_provider_model.id,
             name=llm_provider_model.name,
@@ -202,10 +150,6 @@ class LLMProviderView(LLMProvider):
             api_base=llm_provider_model.api_base,
             api_version=llm_provider_model.api_version,
             custom_config=llm_provider_model.custom_config,
-            default_model_name=default_model_name,
-            is_default_provider=is_default_provider,
-            is_default_vision_provider=is_default_vision_provider,
-            default_vision_model=default_vision_model,
             is_public=llm_provider_model.is_public,
             is_auto_mode=llm_provider_model.is_auto_mode,
             groups=groups,
@@ -425,3 +369,38 @@ class OpenRouterFinalModelResponse(BaseModel):
         int | None
     )  # From OpenRouter API context_length (may be missing for some models)
     supports_image_input: bool
+
+
+class DefaultModel(BaseModel):
+    provider_id: int
+    model_name: str
+
+    @classmethod
+    def from_model_config(
+        cls, model_config: ModelConfigurationModel | None
+    ) -> DefaultModel | None:
+        if not model_config:
+            return None
+        return cls(
+            provider_id=model_config.llm_provider_id,
+            model_name=model_config.name,
+        )
+
+
+class LLMProviderResponse(BaseModel, Generic[T]):
+    providers: list[T]
+    default_text: DefaultModel | None = None
+    default_vision: DefaultModel | None = None
+
+    @classmethod
+    def from_models(
+        cls,
+        providers: list[T],
+        default_text: DefaultModel | None = None,
+        default_vision: DefaultModel | None = None,
+    ) -> LLMProviderResponse[T]:
+        return cls(
+            providers=providers,
+            default_text=default_text,
+            default_vision=default_vision,
+        )

backend/onyx/server/settings/models.py

@@ -19,6 +19,7 @@ class ApplicationStatus(str, Enum):
     PAYMENT_REMINDER = "payment_reminder"
     GRACE_PERIOD = "grace_period"
     GATED_ACCESS = "gated_access"
+    SEAT_LIMIT_EXCEEDED = "seat_limit_exceeded"
 
 
 class Notification(BaseModel):
@@ -82,6 +83,10 @@ class Settings(BaseModel):
     # Default Assistant settings
     disable_default_assistant: bool | None = False
 
+    # Seat usage - populated by license enforcement when seat limit is exceeded
+    seat_count: int | None = None
+    used_seats: int | None = None
+
     # OpenSearch migration
     opensearch_indexing_enabled: bool = False
 

backend/onyx/setup.py

@@ -25,6 +25,7 @@ from onyx.db.enums import EmbeddingPrecision
 from onyx.db.index_attempt import cancel_indexing_attempts_past_model
 from onyx.db.index_attempt import expire_index_attempts
 from onyx.db.llm import fetch_default_llm_model
+from onyx.db.llm import fetch_existing_llm_provider
 from onyx.db.llm import update_default_provider
 from onyx.db.llm import upsert_llm_provider
 from onyx.db.search_settings import get_active_search_settings
@@ -254,14 +255,18 @@ def setup_postgres(db_session: Session) -> None:
         logger.notice("Setting up default OpenAI LLM for dev.")
 
         llm_model = GEN_AI_MODEL_VERSION or "gpt-4o-mini"
+        provider_name = "DevEnvPresetOpenAI"
+        existing = fetch_existing_llm_provider(
+            name=provider_name, db_session=db_session
+        )
         model_req = LLMProviderUpsertRequest(
-            name="DevEnvPresetOpenAI",
+            id=existing.id if existing else None,
+            name=provider_name,
             provider=LlmProviderNames.OPENAI,
             api_key=GEN_AI_API_KEY,
             api_base=None,
             api_version=None,
             custom_config=None,
-            default_model_name=llm_model,
             is_public=True,
             groups=[],
             model_configurations=[
@@ -273,7 +278,9 @@ def setup_postgres(db_session: Session) -> None:
         new_llm_provider = upsert_llm_provider(
             llm_provider_upsert_request=model_req, db_session=db_session
         )
-        update_default_provider(provider_id=new_llm_provider.id, db_session=db_session)
+        update_default_provider(
+            provider_id=new_llm_provider.id, model_name=llm_model, db_session=db_session
+        )
 
 
 def update_default_multipass_indexing(db_session: Session) -> None:

backend/tests/daily/llm/test_bedrock.py

@@ -17,7 +17,7 @@ def test_bedrock_llm_configuration(client: TestClient) -> None:
     # Prepare the test request payload
     test_request: dict[str, Any] = {
         "provider": LlmProviderNames.BEDROCK,
-        "default_model_name": _DEFAULT_BEDROCK_MODEL,
+        "model": _DEFAULT_BEDROCK_MODEL,
         "api_key": None,
         "api_base": None,
         "api_version": None,
@@ -44,7 +44,7 @@ def test_bedrock_llm_configuration_invalid_key(client: TestClient) -> None:
     # Prepare the test request payload with invalid credentials
     test_request: dict[str, Any] = {
         "provider": LlmProviderNames.BEDROCK,
-        "default_model_name": _DEFAULT_BEDROCK_MODEL,
+        "model": _DEFAULT_BEDROCK_MODEL,
         "api_key": None,
         "api_base": None,
         "api_version": None,

backend/tests/external_dependency_unit/answer/conftest.py

@@ -28,7 +28,6 @@ def ensure_default_llm_provider(db_session: Session) -> None:
             provider=LlmProviderNames.OPENAI,
             api_key=os.environ.get("OPENAI_API_KEY", "test"),
             is_public=True,
-            default_model_name="gpt-4o-mini",
             model_configurations=[
                 ModelConfigurationUpsertRequest(
                     name="gpt-4o-mini",
@@ -41,7 +40,7 @@ def ensure_default_llm_provider(db_session: Session) -> None:
             llm_provider_upsert_request=llm_provider_request,
             db_session=db_session,
         )
-        update_default_provider(provider.id, db_session)
+        update_default_provider(provider.id, "gpt-4o-mini", db_session)
     except Exception as exc:  # pragma: no cover - only hits on duplicate setup issues
         # Rollback to clear the pending transaction state
         db_session.rollback()

backend/tests/external_dependency_unit/answer/test_answer_without_openai.py

@@ -47,7 +47,6 @@ def test_answer_with_only_anthropic_provider(
             name=provider_name,
             provider=LlmProviderNames.ANTHROPIC,
             api_key=anthropic_api_key,
-            default_model_name=anthropic_model,
             is_public=True,
             groups=[],
             model_configurations=[
@@ -59,7 +58,7 @@ def test_answer_with_only_anthropic_provider(
     )
 
     try:
-        update_default_provider(anthropic_provider.id, db_session)
+        update_default_provider(anthropic_provider.id, anthropic_model, db_session)
 
         test_user = create_test_user(db_session, email_prefix="anthropic_only")
         chat_session = create_chat_session(

backend/tests/external_dependency_unit/llm/test_llm_provider.py

@@ -29,6 +29,7 @@ from onyx.server.manage.llm.api import (
     test_llm_configuration as run_test_llm_configuration,
 )
 from onyx.server.manage.llm.models import LLMProviderUpsertRequest
+from onyx.server.manage.llm.models import LLMProviderView
 from onyx.server.manage.llm.models import ModelConfigurationUpsertRequest
 from onyx.server.manage.llm.models import TestLLMRequest as LLMTestRequest
 
@@ -44,15 +45,14 @@ def _create_test_provider(
     db_session: Session,
     name: str,
     api_key: str = "sk-test-key-00000000000000000000000000000000000",
-) -> None:
+) -> LLMProviderView:
     """Helper to create a test LLM provider in the database."""
-    upsert_llm_provider(
+    return upsert_llm_provider(
         LLMProviderUpsertRequest(
             name=name,
             provider=LlmProviderNames.OPENAI,
             api_key=api_key,
             api_key_changed=True,
-            default_model_name="gpt-4o-mini",
             model_configurations=[
                 ModelConfigurationUpsertRequest(name="gpt-4o-mini", is_visible=True)
             ],
@@ -102,17 +102,11 @@ class TestLLMConfigurationEndpoint:
                 # This should complete without exception
                 run_test_llm_configuration(
                     test_llm_request=LLMTestRequest(
-                        name=None,  # New provider (not in DB)
                         provider=LlmProviderNames.OPENAI,
                         api_key="sk-new-test-key-0000000000000000000000000000",
                         api_key_changed=True,
                         custom_config_changed=False,
-                        default_model_name="gpt-4o-mini",
-                        model_configurations=[
-                            ModelConfigurationUpsertRequest(
-                                name="gpt-4o-mini", is_visible=True
-                            )
-                        ],
+                        model="gpt-4o-mini",
                     ),
                     _=_create_mock_admin(),
                     db_session=db_session,
@@ -152,17 +146,11 @@ class TestLLMConfigurationEndpoint:
                 with pytest.raises(HTTPException) as exc_info:
                     run_test_llm_configuration(
                         test_llm_request=LLMTestRequest(
-                            name=None,
                             provider=LlmProviderNames.OPENAI,
                             api_key="sk-invalid-key-00000000000000000000000000",
                             api_key_changed=True,
                             custom_config_changed=False,
-                            default_model_name="gpt-4o-mini",
-                            model_configurations=[
-                                ModelConfigurationUpsertRequest(
-                                    name="gpt-4o-mini", is_visible=True
-                                )
-                            ],
+                            model="gpt-4o-mini",
                         ),
                         _=_create_mock_admin(),
                         db_session=db_session,
@@ -194,7 +182,9 @@ class TestLLMConfigurationEndpoint:
 
         try:
             # First, create the provider in the database
-            _create_test_provider(db_session, provider_name, api_key=original_api_key)
+            provider = _create_test_provider(
+                db_session, provider_name, api_key=original_api_key
+            )
 
             with patch(
                 "onyx.server.manage.llm.api.test_llm", side_effect=mock_test_llm_capture
@@ -202,17 +192,12 @@ class TestLLMConfigurationEndpoint:
                 # Test with api_key_changed=False - should use stored key
                 run_test_llm_configuration(
                     test_llm_request=LLMTestRequest(
-                        name=provider_name,  # Existing provider
+                        id=provider.id,
                         provider=LlmProviderNames.OPENAI,
                         api_key=None,  # Not providing a new key
                         api_key_changed=False,  # Using existing key
                         custom_config_changed=False,
-                        default_model_name="gpt-4o-mini",
-                        model_configurations=[
-                            ModelConfigurationUpsertRequest(
-                                name="gpt-4o-mini", is_visible=True
-                            )
-                        ],
+                        model="gpt-4o-mini",
                     ),
                     _=_create_mock_admin(),
                     db_session=db_session,
@@ -246,7 +231,9 @@ class TestLLMConfigurationEndpoint:
 
         try:
             # First, create the provider in the database
-            _create_test_provider(db_session, provider_name, api_key=original_api_key)
+            provider = _create_test_provider(
+                db_session, provider_name, api_key=original_api_key
+            )
 
             with patch(
                 "onyx.server.manage.llm.api.test_llm", side_effect=mock_test_llm_capture
@@ -254,17 +241,12 @@ class TestLLMConfigurationEndpoint:
                 # Test with api_key_changed=True - should use new key
                 run_test_llm_configuration(
                     test_llm_request=LLMTestRequest(
-                        name=provider_name,  # Existing provider
+                        id=provider.id,
                         provider=LlmProviderNames.OPENAI,
                         api_key=new_api_key,  # Providing a new key
                         api_key_changed=True,  # Key is being changed
                         custom_config_changed=False,
-                        default_model_name="gpt-4o-mini",
-                        model_configurations=[
-                            ModelConfigurationUpsertRequest(
-                                name="gpt-4o-mini", is_visible=True
-                            )
-                        ],
+                        model="gpt-4o-mini",
                     ),
                     _=_create_mock_admin(),
                     db_session=db_session,
@@ -297,7 +279,7 @@ class TestLLMConfigurationEndpoint:
 
         try:
             # First, create the provider in the database with custom_config
-            upsert_llm_provider(
+            provider = upsert_llm_provider(
                 LLMProviderUpsertRequest(
                     name=provider_name,
                     provider=LlmProviderNames.OPENAI,
@@ -305,7 +287,6 @@ class TestLLMConfigurationEndpoint:
                     api_key_changed=True,
                     custom_config=original_custom_config,
                     custom_config_changed=True,
-                    default_model_name="gpt-4o-mini",
                     model_configurations=[
                         ModelConfigurationUpsertRequest(
                             name="gpt-4o-mini", is_visible=True
@@ -321,18 +302,13 @@ class TestLLMConfigurationEndpoint:
                 # Test with custom_config_changed=False - should use stored config
                 run_test_llm_configuration(
                     test_llm_request=LLMTestRequest(
-                        name=provider_name,
+                        id=provider.id,
                         provider=LlmProviderNames.OPENAI,
                         api_key=None,
                         api_key_changed=False,
                         custom_config=None,  # Not providing new config
                         custom_config_changed=False,  # Using existing config
-                        default_model_name="gpt-4o-mini",
-                        model_configurations=[
-                            ModelConfigurationUpsertRequest(
-                                name="gpt-4o-mini", is_visible=True
-                            )
-                        ],
+                        model="gpt-4o-mini",
                     ),
                     _=_create_mock_admin(),
                     db_session=db_session,
@@ -368,17 +344,11 @@ class TestLLMConfigurationEndpoint:
                 for model_name in test_models:
                     run_test_llm_configuration(
                         test_llm_request=LLMTestRequest(
-                            name=None,
                             provider=LlmProviderNames.OPENAI,
                             api_key="sk-test-key-00000000000000000000000000000000000",
                             api_key_changed=True,
                             custom_config_changed=False,
-                            default_model_name=model_name,
-                            model_configurations=[
-                                ModelConfigurationUpsertRequest(
-                                    name=model_name, is_visible=True
-                                )
-                            ],
+                            model=model_name,
                         ),
                         _=_create_mock_admin(),
                         db_session=db_session,
@@ -442,7 +412,6 @@ class TestDefaultProviderEndpoint:
                     provider=LlmProviderNames.OPENAI,
                     api_key=provider_1_api_key,
                     api_key_changed=True,
-                    default_model_name=provider_1_initial_model,
                     model_configurations=[
                         ModelConfigurationUpsertRequest(name="gpt-4", is_visible=True),
                         ModelConfigurationUpsertRequest(name="gpt-4o", is_visible=True),
@@ -452,7 +421,7 @@ class TestDefaultProviderEndpoint:
             )
 
             # Set provider 1 as the default provider explicitly
-            update_default_provider(provider_1.id, db_session)
+            update_default_provider(provider_1.id, provider_1_initial_model, db_session)
 
             # Step 2: Call run_test_default_provider - should use provider 1's default model
             with patch(
@@ -472,7 +441,6 @@ class TestDefaultProviderEndpoint:
                     provider=LlmProviderNames.OPENAI,
                     api_key=provider_2_api_key,
                     api_key_changed=True,
-                    default_model_name=provider_2_default_model,
                     model_configurations=[
                         ModelConfigurationUpsertRequest(
                             name="gpt-4o-mini", is_visible=True
@@ -499,11 +467,11 @@ class TestDefaultProviderEndpoint:
             # Step 5: Update provider 1's default model
             upsert_llm_provider(
                 LLMProviderUpsertRequest(
+                    id=provider_1.id,
                     name=provider_1_name,
                     provider=LlmProviderNames.OPENAI,
                     api_key=provider_1_api_key,
                     api_key_changed=True,
-                    default_model_name=provider_1_updated_model,  # Changed
                     model_configurations=[
                         ModelConfigurationUpsertRequest(name="gpt-4", is_visible=True),
                         ModelConfigurationUpsertRequest(name="gpt-4o", is_visible=True),
@@ -512,6 +480,9 @@ class TestDefaultProviderEndpoint:
                 db_session=db_session,
             )
 
+            # Set provider 1's default model to the updated model
+            update_default_provider(provider_1.id, provider_1_updated_model, db_session)
+
             # Step 6: Call run_test_default_provider - should use new model on provider 1
             with patch(
                 "onyx.server.manage.llm.api.test_llm", side_effect=mock_test_llm_capture
@@ -524,7 +495,7 @@ class TestDefaultProviderEndpoint:
             captured_llms.clear()
 
             # Step 7: Change the default provider to provider 2
-            update_default_provider(provider_2.id, db_session)
+            update_default_provider(provider_2.id, provider_2_default_model, db_session)
 
             # Step 8: Call run_test_default_provider - should use provider 2
             with patch(
@@ -596,7 +567,6 @@ class TestDefaultProviderEndpoint:
                     provider=LlmProviderNames.OPENAI,
                     api_key="sk-test-key-00000000000000000000000000000000000",
                     api_key_changed=True,
-                    default_model_name="gpt-4o-mini",
                     model_configurations=[
                         ModelConfigurationUpsertRequest(
                             name="gpt-4o-mini", is_visible=True
@@ -605,7 +575,7 @@ class TestDefaultProviderEndpoint:
                 ),
                 db_session=db_session,
             )
-            update_default_provider(provider.id, db_session)
+            update_default_provider(provider.id, "gpt-4o-mini", db_session)
 
             # Test should fail
             with patch(

backend/tests/external_dependency_unit/llm/test_llm_provider_api_base.py

@@ -49,7 +49,6 @@ def _create_test_provider(
             api_key_changed=True,
             api_base=api_base,
             custom_config=custom_config,
-            default_model_name="gpt-4o-mini",
             model_configurations=[
                 ModelConfigurationUpsertRequest(name="gpt-4o-mini", is_visible=True)
             ],
@@ -91,14 +90,14 @@ class TestLLMProviderChanges:
         the API key should be blocked.
         """
         try:
-            _create_test_provider(db_session, provider_name)
+            provider = _create_test_provider(db_session, provider_name)
 
             with patch("onyx.server.manage.llm.api.MULTI_TENANT", True):
                 update_request = LLMProviderUpsertRequest(
+                    id=provider.id,
                     name=provider_name,
                     provider=LlmProviderNames.OPENAI,
                     api_base="https://attacker.example.com",
-                    default_model_name="gpt-4o-mini",
                 )
 
                 with pytest.raises(HTTPException) as exc_info:
@@ -125,16 +124,16 @@ class TestLLMProviderChanges:
         Changing api_base IS allowed when the API key is also being changed.
         """
         try:
-            _create_test_provider(db_session, provider_name)
+            provider = _create_test_provider(db_session, provider_name)
 
             with patch("onyx.server.manage.llm.api.MULTI_TENANT", True):
                 update_request = LLMProviderUpsertRequest(
+                    id=provider.id,
                     name=provider_name,
                     provider=LlmProviderNames.OPENAI,
                     api_key="sk-new-key-00000000000000000000000000000000000",
                     api_key_changed=True,
                     api_base="https://custom-endpoint.example.com/v1",
-                    default_model_name="gpt-4o-mini",
                 )
 
                 result = put_llm_provider(
@@ -159,14 +158,16 @@ class TestLLMProviderChanges:
         original_api_base = "https://original.example.com/v1"
 
         try:
-            _create_test_provider(db_session, provider_name, api_base=original_api_base)
+            provider = _create_test_provider(
+                db_session, provider_name, api_base=original_api_base
+            )
 
             with patch("onyx.server.manage.llm.api.MULTI_TENANT", True):
                 update_request = LLMProviderUpsertRequest(
+                    id=provider.id,
                     name=provider_name,
                     provider=LlmProviderNames.OPENAI,
                     api_base=original_api_base,
-                    default_model_name="gpt-4o-mini",
                 )
 
                 result = put_llm_provider(
@@ -190,14 +191,14 @@ class TestLLMProviderChanges:
         changes. This allows model-only updates when provider has no custom base URL.
         """
         try:
-            _create_test_provider(db_session, provider_name, api_base=None)
+            view = _create_test_provider(db_session, provider_name, api_base=None)
 
             with patch("onyx.server.manage.llm.api.MULTI_TENANT", True):
                 update_request = LLMProviderUpsertRequest(
+                    id=view.id,
                     name=provider_name,
                     provider=LlmProviderNames.OPENAI,
                     api_base="",
-                    default_model_name="gpt-4o-mini",
                 )
 
                 result = put_llm_provider(
@@ -223,14 +224,16 @@ class TestLLMProviderChanges:
         original_api_base = "https://original.example.com/v1"
 
         try:
-            _create_test_provider(db_session, provider_name, api_base=original_api_base)
+            provider = _create_test_provider(
+                db_session, provider_name, api_base=original_api_base
+            )
 
             with patch("onyx.server.manage.llm.api.MULTI_TENANT", True):
                 update_request = LLMProviderUpsertRequest(
+                    id=provider.id,
                     name=provider_name,
                     provider=LlmProviderNames.OPENAI,
                     api_base=None,
-                    default_model_name="gpt-4o-mini",
                 )
 
                 with pytest.raises(HTTPException) as exc_info:
@@ -259,14 +262,14 @@ class TestLLMProviderChanges:
         users have full control over their deployment.
         """
         try:
-            _create_test_provider(db_session, provider_name)
+            provider = _create_test_provider(db_session, provider_name)
 
             with patch("onyx.server.manage.llm.api.MULTI_TENANT", False):
                 update_request = LLMProviderUpsertRequest(
+                    id=provider.id,
                     name=provider_name,
                     provider=LlmProviderNames.OPENAI,
                     api_base="https://custom.example.com/v1",
-                    default_model_name="gpt-4o-mini",
                 )
 
                 result = put_llm_provider(
@@ -297,7 +300,6 @@ class TestLLMProviderChanges:
                     api_key="sk-new-key-00000000000000000000000000000000000",
                     api_key_changed=True,
                     api_base="https://custom.example.com/v1",
-                    default_model_name="gpt-4o-mini",
                 )
 
                 result = put_llm_provider(
@@ -322,7 +324,7 @@ class TestLLMProviderChanges:
         redirect LLM API requests).
         """
         try:
-            _create_test_provider(
+            provider = _create_test_provider(
                 db_session,
                 provider_name,
                 custom_config={"SOME_CONFIG": "original_value"},
@@ -330,11 +332,11 @@ class TestLLMProviderChanges:
 
             with patch("onyx.server.manage.llm.api.MULTI_TENANT", True):
                 update_request = LLMProviderUpsertRequest(
+                    id=provider.id,
                     name=provider_name,
                     provider=LlmProviderNames.OPENAI,
                     custom_config={"OPENAI_API_BASE": "https://attacker.example.com"},
                     custom_config_changed=True,
-                    default_model_name="gpt-4o-mini",
                 )
 
                 with pytest.raises(HTTPException) as exc_info:
@@ -362,15 +364,15 @@ class TestLLMProviderChanges:
         without changing the API key.
         """
         try:
-            _create_test_provider(db_session, provider_name)
+            provider = _create_test_provider(db_session, provider_name)
 
             with patch("onyx.server.manage.llm.api.MULTI_TENANT", True):
                 update_request = LLMProviderUpsertRequest(
+                    id=provider.id,
                     name=provider_name,
                     provider=LlmProviderNames.OPENAI,
                     custom_config={"OPENAI_API_BASE": "https://attacker.example.com"},
                     custom_config_changed=True,
-                    default_model_name="gpt-4o-mini",
                 )
 
                 with pytest.raises(HTTPException) as exc_info:
@@ -399,7 +401,7 @@ class TestLLMProviderChanges:
         new_config = {"AWS_REGION_NAME": "us-west-2"}
 
         try:
-            _create_test_provider(
+            provider = _create_test_provider(
                 db_session,
                 provider_name,
                 custom_config={"AWS_REGION_NAME": "us-east-1"},
@@ -407,13 +409,13 @@ class TestLLMProviderChanges:
 
             with patch("onyx.server.manage.llm.api.MULTI_TENANT", True):
                 update_request = LLMProviderUpsertRequest(
+                    id=provider.id,
                     name=provider_name,
                     provider=LlmProviderNames.OPENAI,
                     api_key="sk-new-key-00000000000000000000000000000000000",
                     api_key_changed=True,
                     custom_config_changed=True,
                     custom_config=new_config,
-                    default_model_name="gpt-4o-mini",
                 )
 
                 result = put_llm_provider(
@@ -438,17 +440,17 @@ class TestLLMProviderChanges:
         original_config = {"AWS_REGION_NAME": "us-east-1"}
 
         try:
-            _create_test_provider(
+            provider = _create_test_provider(
                 db_session, provider_name, custom_config=original_config
             )
 
             with patch("onyx.server.manage.llm.api.MULTI_TENANT", True):
                 update_request = LLMProviderUpsertRequest(
+                    id=provider.id,
                     name=provider_name,
                     provider=LlmProviderNames.OPENAI,
                     custom_config=original_config,
                     custom_config_changed=True,
-                    default_model_name="gpt-4o-mini",
                 )
 
                 result = put_llm_provider(
@@ -474,7 +476,7 @@ class TestLLMProviderChanges:
         new_config = {"AWS_REGION_NAME": "eu-west-1"}
 
         try:
-            _create_test_provider(
+            provider = _create_test_provider(
                 db_session,
                 provider_name,
                 custom_config={"AWS_REGION_NAME": "us-east-1"},
@@ -482,10 +484,10 @@ class TestLLMProviderChanges:
 
             with patch("onyx.server.manage.llm.api.MULTI_TENANT", False):
                 update_request = LLMProviderUpsertRequest(
+                    id=provider.id,
                     name=provider_name,
                     provider=LlmProviderNames.OPENAI,
                     custom_config=new_config,
-                    default_model_name="gpt-4o-mini",
                     custom_config_changed=True,
                 )
 
@@ -530,14 +532,8 @@ def test_upload_with_custom_config_then_change(
         with patch("onyx.server.manage.llm.api.test_llm", side_effect=capture_test_llm):
             run_llm_config_test(
                 LLMTestRequest(
-                    name=name,
                     provider=provider_name,
-                    default_model_name=default_model_name,
-                    model_configurations=[
-                        ModelConfigurationUpsertRequest(
-                            name=default_model_name, is_visible=True
-                        )
-                    ],
+                    model=default_model_name,
                     api_key_changed=False,
                     custom_config_changed=True,
                     custom_config=custom_config,
@@ -546,11 +542,10 @@ def test_upload_with_custom_config_then_change(
                 db_session=db_session,
             )
 
-            put_llm_provider(
+            provider = put_llm_provider(
                 llm_provider_upsert_request=LLMProviderUpsertRequest(
                     name=name,
                     provider=provider_name,
-                    default_model_name=default_model_name,
                     custom_config=custom_config,
                     model_configurations=[
                         ModelConfigurationUpsertRequest(
@@ -569,14 +564,9 @@ def test_upload_with_custom_config_then_change(
             # Turn auto mode off
             run_llm_config_test(
                 LLMTestRequest(
-                    name=name,
+                    id=provider.id,
                     provider=provider_name,
-                    default_model_name=default_model_name,
-                    model_configurations=[
-                        ModelConfigurationUpsertRequest(
-                            name=default_model_name, is_visible=True
-                        )
-                    ],
+                    model=default_model_name,
                     api_key_changed=False,
                     custom_config_changed=False,
                 ),
@@ -586,9 +576,9 @@ def test_upload_with_custom_config_then_change(
 
             put_llm_provider(
                 llm_provider_upsert_request=LLMProviderUpsertRequest(
+                    id=provider.id,
                     name=name,
                     provider=provider_name,
-                    default_model_name=default_model_name,
                     model_configurations=[
                         ModelConfigurationUpsertRequest(
                             name=default_model_name, is_visible=True
@@ -616,13 +606,13 @@ def test_upload_with_custom_config_then_change(
                 )
 
             # Check inside the database and check that custom_config is the same as the original
-            provider = fetch_existing_llm_provider(name=name, db_session=db_session)
-            if not provider:
+            db_provider = fetch_existing_llm_provider(name=name, db_session=db_session)
+            if not db_provider:
                 assert False, "Provider not found in the database"
 
-            assert provider.custom_config == custom_config, (
+            assert db_provider.custom_config == custom_config, (
                 f"Expected custom_config {custom_config}, "
-                f"but got {provider.custom_config}"
+                f"but got {db_provider.custom_config}"
             )
     finally:
         db_session.rollback()
@@ -642,11 +632,10 @@ def test_preserves_masked_sensitive_custom_config_on_provider_update(
     }
 
     try:
-        put_llm_provider(
+        view = put_llm_provider(
             llm_provider_upsert_request=LLMProviderUpsertRequest(
                 name=name,
                 provider=provider,
-                default_model_name=default_model_name,
                 custom_config=original_custom_config,
                 model_configurations=[
                     ModelConfigurationUpsertRequest(
@@ -665,9 +654,9 @@ def test_preserves_masked_sensitive_custom_config_on_provider_update(
         with patch("onyx.server.manage.llm.api.MULTI_TENANT", False):
             put_llm_provider(
                 llm_provider_upsert_request=LLMProviderUpsertRequest(
+                    id=view.id,
                     name=name,
                     provider=provider,
-                    default_model_name=default_model_name,
                     custom_config={
                         "vertex_credentials": _mask_string(
                             original_custom_config["vertex_credentials"]
@@ -706,7 +695,7 @@ def test_preserves_masked_sensitive_custom_config_on_test_request(
 ) -> None:
     """LLM test should restore masked sensitive custom config values before invocation."""
     name = f"test-provider-vertex-test-{uuid4().hex[:8]}"
-    provider = LlmProviderNames.VERTEX_AI.value
+    provider_name = LlmProviderNames.VERTEX_AI.value
     default_model_name = "gemini-2.5-pro"
     original_custom_config = {
         "vertex_credentials": '{"type":"service_account","private_key":"REAL_PRIVATE_KEY"}',
@@ -719,11 +708,10 @@ def test_preserves_masked_sensitive_custom_config_on_test_request(
         return ""
 
     try:
-        put_llm_provider(
+        provider = put_llm_provider(
             llm_provider_upsert_request=LLMProviderUpsertRequest(
                 name=name,
-                provider=provider,
-                default_model_name=default_model_name,
+                provider=provider_name,
                 custom_config=original_custom_config,
                 model_configurations=[
                     ModelConfigurationUpsertRequest(
@@ -742,14 +730,9 @@ def test_preserves_masked_sensitive_custom_config_on_test_request(
         with patch("onyx.server.manage.llm.api.test_llm", side_effect=capture_test_llm):
             run_llm_config_test(
                 LLMTestRequest(
-                    name=name,
-                    provider=provider,
-                    default_model_name=default_model_name,
-                    model_configurations=[
-                        ModelConfigurationUpsertRequest(
-                            name=default_model_name, is_visible=True
-                        )
-                    ],
+                    id=provider.id,
+                    provider=provider_name,
+                    model=default_model_name,
                     api_key_changed=False,
                     custom_config_changed=True,
                     custom_config={

backend/tests/external_dependency_unit/llm/test_llm_provider_auto_mode.py

@@ -15,9 +15,11 @@ import pytest
 from sqlalchemy.orm import Session
 
 from onyx.db.enums import LLMModelFlowType
+from onyx.db.llm import fetch_auto_mode_providers
 from onyx.db.llm import fetch_default_llm_model
 from onyx.db.llm import fetch_existing_llm_provider
 from onyx.db.llm import fetch_existing_llm_providers
+from onyx.db.llm import fetch_llm_provider_view
 from onyx.db.llm import remove_llm_provider
 from onyx.db.llm import sync_auto_mode_models
 from onyx.db.llm import update_default_provider
@@ -135,7 +137,6 @@ class TestAutoModeSyncFeature:
                         api_key="sk-test-key-00000000000000000000000000000000000",
                         api_key_changed=True,
                         is_auto_mode=True,
-                        default_model_name=expected_default_model,
                         model_configurations=[],  # No model configs provided
                     ),
                     is_creation=True,
@@ -163,13 +164,8 @@ class TestAutoModeSyncFeature:
                 if mc.name in all_expected_models:
                     assert mc.is_visible is True, f"Model '{mc.name}' should be visible"
 
-            # Verify the default model was set correctly
-            assert (
-                provider.default_model_name == expected_default_model
-            ), f"Default model should be '{expected_default_model}'"
-
             # Step 4: Set the provider as default
-            update_default_provider(provider.id, db_session)
+            update_default_provider(provider.id, expected_default_model, db_session)
 
             # Step 5: Fetch the default provider and verify
             default_model = fetch_default_llm_model(db_session)
@@ -238,7 +234,6 @@ class TestAutoModeSyncFeature:
                         api_key="sk-test-key-00000000000000000000000000000000000",
                         api_key_changed=True,
                         is_auto_mode=True,
-                        default_model_name="gpt-4o",
                         model_configurations=[],
                     ),
                     is_creation=True,
@@ -317,7 +312,6 @@ class TestAutoModeSyncFeature:
                     api_key="sk-test-key-00000000000000000000000000000000000",
                     api_key_changed=True,
                     is_auto_mode=False,  # Not in auto mode initially
-                    default_model_name="gpt-4",
                     model_configurations=initial_models,
                 ),
                 is_creation=True,
@@ -326,13 +320,13 @@ class TestAutoModeSyncFeature:
             )
 
             # Verify initial state: all models are visible
-            provider = fetch_existing_llm_provider(
+            initial_provider = fetch_existing_llm_provider(
                 name=provider_name, db_session=db_session
             )
-            assert provider is not None
-            assert provider.is_auto_mode is False
+            assert initial_provider is not None
+            assert initial_provider.is_auto_mode is False
 
-            for mc in provider.model_configurations:
+            for mc in initial_provider.model_configurations:
                 assert (
                     mc.is_visible is True
                 ), f"Initial model '{mc.name}' should be visible"
@@ -344,12 +338,12 @@ class TestAutoModeSyncFeature:
             ):
                 put_llm_provider(
                     llm_provider_upsert_request=LLMProviderUpsertRequest(
+                        id=initial_provider.id,
                         name=provider_name,
                         provider=LlmProviderNames.OPENAI,
                         api_key=None,  # Not changing API key
                         api_key_changed=False,
                         is_auto_mode=True,  # Now enabling auto mode
-                        default_model_name=auto_mode_default,
                         model_configurations=[],  # Auto mode will sync from config
                     ),
                     is_creation=False,  # This is an update
@@ -360,15 +354,15 @@ class TestAutoModeSyncFeature:
             # Step 3: Verify model visibility after auto mode transition
             # Expire session cache to force fresh fetch after sync_auto_mode_models committed
             db_session.expire_all()
-            provider = fetch_existing_llm_provider(
-                name=provider_name, db_session=db_session
+            provider_view = fetch_llm_provider_view(
+                provider_name=provider_name, db_session=db_session
             )
-            assert provider is not None
-            assert provider.is_auto_mode is True
+            assert provider_view is not None
+            assert provider_view.is_auto_mode is True
 
             # Build a map of model name -> visibility
             model_visibility = {
-                mc.name: mc.is_visible for mc in provider.model_configurations
+                mc.name: mc.is_visible for mc in provider_view.model_configurations
             }
 
             # Models in auto mode config should be visible
@@ -388,9 +382,6 @@ class TestAutoModeSyncFeature:
                         model_visibility[model_name] is False
                     ), f"Model '{model_name}' not in auto config should NOT be visible"
 
-            # Verify the default model was updated
-            assert provider.default_model_name == auto_mode_default
-
         finally:
             db_session.rollback()
             _cleanup_provider(db_session, provider_name)
@@ -432,8 +423,12 @@ class TestAutoModeSyncFeature:
                         api_key="sk-test-key-00000000000000000000000000000000000",
                         api_key_changed=True,
                         is_auto_mode=True,
-                        default_model_name="gpt-4o",
-                        model_configurations=[],
+                        model_configurations=[
+                            ModelConfigurationUpsertRequest(
+                                name="gpt-4o",
+                                is_visible=True,
+                            )
+                        ],
                     ),
                     is_creation=True,
                     _=_create_mock_admin(),
@@ -535,7 +530,6 @@ class TestAutoModeSyncFeature:
                         api_key=provider_1_api_key,
                         api_key_changed=True,
                         is_auto_mode=True,
-                        default_model_name=provider_1_default_model,
                         model_configurations=[],
                     ),
                     is_creation=True,
@@ -549,7 +543,7 @@ class TestAutoModeSyncFeature:
                 name=provider_1_name, db_session=db_session
             )
             assert provider_1 is not None
-            update_default_provider(provider_1.id, db_session)
+            update_default_provider(provider_1.id, provider_1_default_model, db_session)
 
             with patch(
                 "onyx.server.manage.llm.api.fetch_llm_recommendations_from_github",
@@ -563,7 +557,6 @@ class TestAutoModeSyncFeature:
                         api_key=provider_2_api_key,
                         api_key_changed=True,
                         is_auto_mode=True,
-                        default_model_name=provider_2_default_model,
                         model_configurations=[],
                     ),
                     is_creation=True,
@@ -584,7 +577,7 @@ class TestAutoModeSyncFeature:
                 name=provider_2_name, db_session=db_session
             )
             assert provider_2 is not None
-            update_default_provider(provider_2.id, db_session)
+            update_default_provider(provider_2.id, provider_2_default_model, db_session)
 
             # Step 5: Verify provider 2 is now the default
             db_session.expire_all()
@@ -644,7 +637,6 @@ class TestAutoModeMissingFlows:
                     api_key="sk-test-key-00000000000000000000000000000000000",
                     api_key_changed=True,
                     is_auto_mode=True,
-                    default_model_name="gpt-4o",
                     model_configurations=[],
                 ),
                 is_creation=True,
@@ -701,3 +693,364 @@ class TestAutoModeMissingFlows:
         finally:
             db_session.rollback()
             _cleanup_provider(db_session, provider_name)
+
+
+class TestAutoModeTransitionsAndResync:
+    """Tests for auto/manual transitions, config evolution, and sync idempotency."""
+
+    def test_auto_to_manual_mode_preserves_models_and_stops_syncing(
+        self,
+        db_session: Session,
+        provider_name: str,
+    ) -> None:
+        """Disabling auto mode should preserve the current model list and
+        prevent future syncs from altering visibility.
+
+        Steps:
+        1. Create provider in auto mode — models synced from config.
+        2. Update provider to manual mode (is_auto_mode=False).
+        3. Verify all models remain with unchanged visibility.
+        4. Call sync_auto_mode_models with a *different* config.
+        5. Verify fetch_auto_mode_providers excludes this provider, so the
+           periodic task would never call sync on it.
+        """
+        initial_config = _create_mock_llm_recommendations(
+            provider=LlmProviderNames.OPENAI,
+            default_model_name="gpt-4o",
+            additional_models=["gpt-4o-mini"],
+        )
+
+        try:
+            # Step 1: Create in auto mode
+            with patch(
+                "onyx.server.manage.llm.api.fetch_llm_recommendations_from_github",
+                return_value=initial_config,
+            ):
+                put_llm_provider(
+                    llm_provider_upsert_request=LLMProviderUpsertRequest(
+                        name=provider_name,
+                        provider=LlmProviderNames.OPENAI,
+                        api_key="sk-test-key-00000000000000000000000000000000000",
+                        api_key_changed=True,
+                        is_auto_mode=True,
+                        model_configurations=[],
+                    ),
+                    is_creation=True,
+                    _=_create_mock_admin(),
+                    db_session=db_session,
+                )
+
+            db_session.expire_all()
+            provider = fetch_existing_llm_provider(
+                name=provider_name, db_session=db_session
+            )
+            assert provider is not None
+            visibility_before = {
+                mc.name: mc.is_visible for mc in provider.model_configurations
+            }
+            assert visibility_before == {"gpt-4o": True, "gpt-4o-mini": True}
+
+            # Step 2: Switch to manual mode
+            put_llm_provider(
+                llm_provider_upsert_request=LLMProviderUpsertRequest(
+                    id=provider.id,
+                    name=provider_name,
+                    provider=LlmProviderNames.OPENAI,
+                    api_key=None,
+                    api_key_changed=False,
+                    is_auto_mode=False,
+                    model_configurations=[
+                        ModelConfigurationUpsertRequest(name="gpt-4o", is_visible=True),
+                        ModelConfigurationUpsertRequest(
+                            name="gpt-4o-mini", is_visible=True
+                        ),
+                    ],
+                ),
+                is_creation=False,
+                _=_create_mock_admin(),
+                db_session=db_session,
+            )
+
+            # Step 3: Models unchanged
+            db_session.expire_all()
+            provider = fetch_existing_llm_provider(
+                name=provider_name, db_session=db_session
+            )
+            assert provider is not None
+            assert provider.is_auto_mode is False
+            visibility_after = {
+                mc.name: mc.is_visible for mc in provider.model_configurations
+            }
+            assert visibility_after == visibility_before
+
+            # Step 4-5: Provider excluded from auto mode queries
+            auto_providers = fetch_auto_mode_providers(db_session)
+            auto_provider_ids = {p.id for p in auto_providers}
+            assert provider.id not in auto_provider_ids
+
+        finally:
+            db_session.rollback()
+            _cleanup_provider(db_session, provider_name)
+
+    def test_resync_adds_new_and_hides_removed_models(
+        self,
+        db_session: Session,
+        provider_name: str,
+    ) -> None:
+        """When the GitHub config changes between syncs, a subsequent sync
+        should add newly listed models and hide models that were removed.
+
+        Steps:
+        1. Create provider in auto mode with config v1: [gpt-4o, gpt-4o-mini].
+        2. Sync with config v2: [gpt-4o, gpt-4-turbo] (gpt-4o-mini removed,
+           gpt-4-turbo added).
+        3. Verify gpt-4o still visible, gpt-4o-mini hidden, gpt-4-turbo added
+           and visible.
+        """
+        config_v1 = _create_mock_llm_recommendations(
+            provider=LlmProviderNames.OPENAI,
+            default_model_name="gpt-4o",
+            additional_models=["gpt-4o-mini"],
+        )
+        config_v2 = _create_mock_llm_recommendations(
+            provider=LlmProviderNames.OPENAI,
+            default_model_name="gpt-4o",
+            additional_models=["gpt-4-turbo"],
+        )
+
+        try:
+            # Step 1: Create with config v1
+            with patch(
+                "onyx.server.manage.llm.api.fetch_llm_recommendations_from_github",
+                return_value=config_v1,
+            ):
+                put_llm_provider(
+                    llm_provider_upsert_request=LLMProviderUpsertRequest(
+                        name=provider_name,
+                        provider=LlmProviderNames.OPENAI,
+                        api_key="sk-test-key-00000000000000000000000000000000000",
+                        api_key_changed=True,
+                        is_auto_mode=True,
+                        model_configurations=[],
+                    ),
+                    is_creation=True,
+                    _=_create_mock_admin(),
+                    db_session=db_session,
+                )
+
+            # Step 2: Re-sync with config v2
+            db_session.expire_all()
+            provider = fetch_existing_llm_provider(
+                name=provider_name, db_session=db_session
+            )
+            assert provider is not None
+
+            changes = sync_auto_mode_models(
+                db_session=db_session,
+                provider=provider,
+                llm_recommendations=config_v2,
+            )
+            assert changes > 0
+
+            # Step 3: Verify
+            db_session.expire_all()
+            provider = fetch_existing_llm_provider(
+                name=provider_name, db_session=db_session
+            )
+            assert provider is not None
+
+            visibility = {
+                mc.name: mc.is_visible for mc in provider.model_configurations
+            }
+
+            # gpt-4o: still in config -> visible
+            assert visibility["gpt-4o"] is True
+            # gpt-4o-mini: removed from config -> hidden (not deleted)
+            assert "gpt-4o-mini" in visibility, "Removed model should still exist in DB"
+            assert visibility["gpt-4o-mini"] is False
+            # gpt-4-turbo: newly added -> visible
+            assert visibility["gpt-4-turbo"] is True
+
+        finally:
+            db_session.rollback()
+            _cleanup_provider(db_session, provider_name)
+
+    def test_sync_is_idempotent(
+        self,
+        db_session: Session,
+        provider_name: str,
+    ) -> None:
+        """Running sync twice with the same config should produce zero
+        changes on the second call."""
+        config = _create_mock_llm_recommendations(
+            provider=LlmProviderNames.OPENAI,
+            default_model_name="gpt-4o",
+            additional_models=["gpt-4o-mini", "gpt-4-turbo"],
+        )
+
+        try:
+            with patch(
+                "onyx.server.manage.llm.api.fetch_llm_recommendations_from_github",
+                return_value=config,
+            ):
+                put_llm_provider(
+                    llm_provider_upsert_request=LLMProviderUpsertRequest(
+                        name=provider_name,
+                        provider=LlmProviderNames.OPENAI,
+                        api_key="sk-test-key-00000000000000000000000000000000000",
+                        api_key_changed=True,
+                        is_auto_mode=True,
+                        model_configurations=[],
+                    ),
+                    is_creation=True,
+                    _=_create_mock_admin(),
+                    db_session=db_session,
+                )
+
+            db_session.expire_all()
+            provider = fetch_existing_llm_provider(
+                name=provider_name, db_session=db_session
+            )
+            assert provider is not None
+
+            # First explicit sync (may report changes if creation already synced)
+            sync_auto_mode_models(
+                db_session=db_session,
+                provider=provider,
+                llm_recommendations=config,
+            )
+
+            # Snapshot state after first sync
+            db_session.expire_all()
+            provider = fetch_existing_llm_provider(
+                name=provider_name, db_session=db_session
+            )
+            assert provider is not None
+            snapshot = {
+                mc.name: (mc.is_visible, mc.display_name)
+                for mc in provider.model_configurations
+            }
+
+            # Second sync — should be a no-op
+            changes = sync_auto_mode_models(
+                db_session=db_session,
+                provider=provider,
+                llm_recommendations=config,
+            )
+            assert (
+                changes == 0
+            ), f"Expected 0 changes on idempotent re-sync, got {changes}"
+
+            # State should be identical
+            db_session.expire_all()
+            provider = fetch_existing_llm_provider(
+                name=provider_name, db_session=db_session
+            )
+            assert provider is not None
+            current = {
+                mc.name: (mc.is_visible, mc.display_name)
+                for mc in provider.model_configurations
+            }
+            assert current == snapshot
+
+        finally:
+            db_session.rollback()
+            _cleanup_provider(db_session, provider_name)
+
+    def test_default_model_hidden_when_removed_from_config(
+        self,
+        db_session: Session,
+        provider_name: str,
+    ) -> None:
+        """When the current default model is removed from the config, sync
+        should hide it. The default model flow row should still exist (it
+        points at the ModelConfiguration), but the model is no longer visible.
+
+        Steps:
+        1. Create provider with config: default=gpt-4o, additional=[gpt-4o-mini].
+        2. Set gpt-4o as the global default.
+        3. Re-sync with config: default=gpt-4o-mini (gpt-4o removed entirely).
+        4. Verify gpt-4o is hidden, gpt-4o-mini is visible, and
+           fetch_default_llm_model still returns a result (the flow row persists).
+        """
+        config_v1 = _create_mock_llm_recommendations(
+            provider=LlmProviderNames.OPENAI,
+            default_model_name="gpt-4o",
+            additional_models=["gpt-4o-mini"],
+        )
+        config_v2 = _create_mock_llm_recommendations(
+            provider=LlmProviderNames.OPENAI,
+            default_model_name="gpt-4o-mini",
+            additional_models=[],
+        )
+
+        try:
+            with patch(
+                "onyx.server.manage.llm.api.fetch_llm_recommendations_from_github",
+                return_value=config_v1,
+            ):
+                put_llm_provider(
+                    llm_provider_upsert_request=LLMProviderUpsertRequest(
+                        name=provider_name,
+                        provider=LlmProviderNames.OPENAI,
+                        api_key="sk-test-key-00000000000000000000000000000000000",
+                        api_key_changed=True,
+                        is_auto_mode=True,
+                        model_configurations=[],
+                    ),
+                    is_creation=True,
+                    _=_create_mock_admin(),
+                    db_session=db_session,
+                )
+
+            # Step 2: Set gpt-4o as global default
+            db_session.expire_all()
+            provider = fetch_existing_llm_provider(
+                name=provider_name, db_session=db_session
+            )
+            assert provider is not None
+            update_default_provider(provider.id, "gpt-4o", db_session)
+
+            default_before = fetch_default_llm_model(db_session)
+            assert default_before is not None
+            assert default_before.name == "gpt-4o"
+
+            # Step 3: Re-sync with config v2 (gpt-4o removed)
+            db_session.expire_all()
+            provider = fetch_existing_llm_provider(
+                name=provider_name, db_session=db_session
+            )
+            assert provider is not None
+
+            changes = sync_auto_mode_models(
+                db_session=db_session,
+                provider=provider,
+                llm_recommendations=config_v2,
+            )
+            assert changes > 0
+
+            # Step 4: Verify visibility
+            db_session.expire_all()
+            provider = fetch_existing_llm_provider(
+                name=provider_name, db_session=db_session
+            )
+            assert provider is not None
+
+            visibility = {
+                mc.name: mc.is_visible for mc in provider.model_configurations
+            }
+            assert visibility["gpt-4o"] is False, "Removed default should be hidden"
+            assert visibility["gpt-4o-mini"] is True, "New default should be visible"
+
+            # The LLMModelFlow row for gpt-4o still exists (is_default=True),
+            # but the model is hidden. fetch_default_llm_model filters on
+            # is_visible=True, so it should NOT return gpt-4o.
+            db_session.expire_all()
+            default_after = fetch_default_llm_model(db_session)
+            assert (
+                default_after is None or default_after.name != "gpt-4o"
+            ), "Hidden model should not be returned as the default"
+
+        finally:
+            db_session.rollback()
+            _cleanup_provider(db_session, provider_name)

backend/tests/external_dependency_unit/llm/test_llm_provider_called.py

@@ -64,7 +64,6 @@ def _create_provider(
             name=name,
             provider=provider,
             api_key="sk-ant-api03-...",
-            default_model_name="claude-3-5-sonnet-20240620",
             is_public=is_public,
             model_configurations=[
                 ModelConfigurationUpsertRequest(
@@ -154,7 +153,9 @@ def test_user_sends_message_to_private_provider(
     )
     _create_provider(db_session, LlmProviderNames.GOOGLE, "private-provider", False)
 
-    update_default_provider(public_provider_id, db_session)
+    update_default_provider(
+        public_provider_id, "claude-3-5-sonnet-20240620", db_session
+    )
 
     try:
         # Create chat session

backend/tests/external_dependency_unit/search_settings/test_search_settings.py

@@ -42,7 +42,6 @@ def _create_llm_provider_and_model(
             name=provider_name,
             provider="openai",
             api_key="test-api-key",
-            default_model_name=model_name,
             model_configurations=[
                 ModelConfigurationUpsertRequest(
                     name=model_name,

backend/tests/external_dependency_unit/slack_bot/test_slack_bot_federated_search.py

@@ -434,7 +434,6 @@ class TestSlackBotFederatedSearch:
                 name=f"test-llm-provider-{uuid4().hex[:8]}",
                 provider=LlmProviderNames.OPENAI,
                 api_key=api_key,
-                default_model_name="gpt-4o",
                 is_public=True,
                 model_configurations=[
                     ModelConfigurationUpsertRequest(
@@ -448,7 +447,7 @@ class TestSlackBotFederatedSearch:
             db_session=db_session,
         )
 
-        update_default_provider(provider_view.id, db_session)
+        update_default_provider(provider_view.id, "gpt-4o", db_session)
 
     def _teardown_common_mocks(self, patches: list) -> None:
         """Stop all patches"""

backend/tests/integration/common_utils/managers/llm_provider.py

@@ -4,10 +4,12 @@ from uuid import uuid4
 import requests
 
 from onyx.llm.constants import LlmProviderNames
+from onyx.server.manage.llm.models import DefaultModel
 from onyx.server.manage.llm.models import LLMProviderUpsertRequest
 from onyx.server.manage.llm.models import LLMProviderView
 from onyx.server.manage.llm.models import ModelConfigurationUpsertRequest
 from tests.integration.common_utils.constants import API_SERVER_URL
+from tests.integration.common_utils.constants import GENERAL_HEADERS
 from tests.integration.common_utils.test_models import DATestLLMProvider
 from tests.integration.common_utils.test_models import DATestUser
 
@@ -32,7 +34,6 @@ class LLMProviderManager:
         llm_provider = LLMProviderUpsertRequest(
             name=name or f"test-provider-{uuid4()}",
             provider=provider or LlmProviderNames.OPENAI,
-            default_model_name=default_model_name or "gpt-4o-mini",
             api_key=api_key or os.environ["OPENAI_API_KEY"],
             api_base=api_base,
             api_version=api_version,
@@ -65,7 +66,7 @@ class LLMProviderManager:
             name=response_data["name"],
             provider=response_data["provider"],
             api_key=response_data["api_key"],
-            default_model_name=response_data["default_model_name"],
+            default_model_name=default_model_name or "gpt-4o-mini",
             is_public=response_data["is_public"],
             is_auto_mode=response_data.get("is_auto_mode", False),
             groups=response_data["groups"],
@@ -75,9 +76,19 @@ class LLMProviderManager:
         )
 
         if set_as_default:
+            if default_model_name is None:
+                default_model_name = "gpt-4o-mini"
             set_default_response = requests.post(
-                f"{API_SERVER_URL}/admin/llm/provider/{llm_response.json()['id']}/default",
-                headers=user_performing_action.headers,
+                f"{API_SERVER_URL}/admin/llm/default",
+                json={
+                    "provider_id": response_data["id"],
+                    "model_name": default_model_name,
+                },
+                headers=(
+                    user_performing_action.headers
+                    if user_performing_action
+                    else GENERAL_HEADERS
+                ),
             )
             set_default_response.raise_for_status()
 
@@ -104,7 +115,7 @@ class LLMProviderManager:
             headers=user_performing_action.headers,
         )
         response.raise_for_status()
-        return [LLMProviderView(**ug) for ug in response.json()]
+        return [LLMProviderView(**p) for p in response.json()["providers"]]
 
     @staticmethod
     def verify(
@@ -113,7 +124,11 @@ class LLMProviderManager:
         verify_deleted: bool = False,
     ) -> None:
         all_llm_providers = LLMProviderManager.get_all(user_performing_action)
+        default_model = LLMProviderManager.get_default_model(user_performing_action)
         for fetched_llm_provider in all_llm_providers:
+            model_names = [
+                model.name for model in fetched_llm_provider.model_configurations
+            ]
             if llm_provider.id == fetched_llm_provider.id:
                 if verify_deleted:
                     raise ValueError(
@@ -126,11 +141,30 @@ class LLMProviderManager:
                 if (
                     fetched_llm_groups == llm_provider_groups
                     and llm_provider.provider == fetched_llm_provider.provider
-                    and llm_provider.default_model_name
-                    == fetched_llm_provider.default_model_name
+                    and (
+                        default_model is None or default_model.model_name in model_names
+                    )
                     and llm_provider.is_public == fetched_llm_provider.is_public
                     and set(fetched_llm_provider.personas) == set(llm_provider.personas)
                 ):
                     return
         if not verify_deleted:
             raise ValueError(f"LLM Provider {llm_provider.id} not found")
+
+    @staticmethod
+    def get_default_model(
+        user_performing_action: DATestUser | None = None,
+    ) -> DefaultModel | None:
+        response = requests.get(
+            f"{API_SERVER_URL}/admin/llm/provider",
+            headers=(
+                user_performing_action.headers
+                if user_performing_action
+                else GENERAL_HEADERS
+            ),
+        )
+        response.raise_for_status()
+        default_text = response.json().get("default_text")
+        if default_text is None:
+            return None
+        return DefaultModel(**default_text)

backend/tests/integration/common_utils/test_models.py

@@ -128,7 +128,7 @@ class DATestLLMProvider(BaseModel):
     name: str
     provider: str
     api_key: str
-    default_model_name: str
+    default_model_name: str | None = None
     is_public: bool
     is_auto_mode: bool = False
     groups: list[int]

backend/tests/integration/tests/llm_auto_update/test_auto_llm_update.py

@@ -42,12 +42,10 @@ def _create_provider_with_api(
     llm_provider_data = {
         "name": name,
         "provider": provider_type,
-        "default_model_name": default_model,
         "api_key": "test-api-key-for-auto-mode-testing",
         "api_base": None,
         "api_version": None,
         "custom_config": None,
-        "fast_default_model_name": default_model,
         "is_public": True,
         "is_auto_mode": is_auto_mode,
         "groups": [],
@@ -72,7 +70,7 @@ def _get_provider_by_id(admin_user: DATestUser, provider_id: int) -> dict:
         headers=admin_user.headers,
     )
     response.raise_for_status()
-    for provider in response.json():
+    for provider in response.json()["providers"]:
         if provider["id"] == provider_id:
             return provider
     raise ValueError(f"Provider with id {provider_id} not found")
@@ -219,15 +217,6 @@ def test_auto_mode_provider_gets_synced_from_github_config(
         "is_visible"
     ], "Outdated model should not be visible after sync"
 
-    # Verify default model was set from GitHub config
-    expected_default = (
-        default_model["name"] if isinstance(default_model, dict) else default_model
-    )
-    assert synced_provider["default_model_name"] == expected_default, (
-        f"Default model should be {expected_default}, "
-        f"got {synced_provider['default_model_name']}"
-    )
-
 
 def test_manual_mode_provider_not_affected_by_auto_sync(
     reset: None,  # noqa: ARG001
@@ -273,7 +262,3 @@ def test_manual_mode_provider_not_affected_by_auto_sync(
         f"Manual mode provider models should not change. "
         f"Initial: {initial_models}, Current: {current_models}"
     )
-
-    assert (
-        updated_provider["default_model_name"] == custom_model
-    ), f"Manual mode default model should remain {custom_model}"

backend/tests/integration/tests/llm_provider/test_llm_provider_access_control.py

@@ -6,20 +6,21 @@ from sqlalchemy.orm import Session
 
 from onyx.context.search.enums import RecencyBiasSetting
 from onyx.db.engine.sql_engine import get_session_with_current_tenant
-from onyx.db.enums import LLMModelFlowType
 from onyx.db.llm import can_user_access_llm_provider
 from onyx.db.llm import fetch_user_group_ids
-from onyx.db.models import LLMModelFlow
+from onyx.db.llm import update_default_provider
+from onyx.db.llm import upsert_llm_provider
 from onyx.db.models import LLMProvider as LLMProviderModel
 from onyx.db.models import LLMProvider__Persona
 from onyx.db.models import LLMProvider__UserGroup
-from onyx.db.models import ModelConfiguration
 from onyx.db.models import Persona
 from onyx.db.models import User
 from onyx.db.models import User__UserGroup
 from onyx.db.models import UserGroup
 from onyx.llm.constants import LlmProviderNames
 from onyx.llm.factory import get_llm_for_persona
+from onyx.server.manage.llm.models import LLMProviderUpsertRequest
+from onyx.server.manage.llm.models import ModelConfigurationUpsertRequest
 from tests.integration.common_utils.constants import API_SERVER_URL
 from tests.integration.common_utils.managers.llm_provider import LLMProviderManager
 from tests.integration.common_utils.managers.persona import PersonaManager
@@ -41,24 +42,30 @@ def _create_llm_provider(
     is_public: bool,
     is_default: bool,
 ) -> LLMProviderModel:
-    provider = LLMProviderModel(
-        name=name,
-        provider=LlmProviderNames.OPENAI,
-        api_key=None,
-        api_base=None,
-        api_version=None,
-        custom_config=None,
-        default_model_name=default_model_name,
-        deployment_name=None,
-        is_public=is_public,
-        # Use None instead of False to avoid unique constraint violation
-        # The is_default_provider column has unique=True, so only one True and one False allowed
-        is_default_provider=is_default if is_default else None,
-        is_default_vision_provider=False,
-        default_vision_model=None,
+    _provider = upsert_llm_provider(
+        llm_provider_upsert_request=LLMProviderUpsertRequest(
+            name=name,
+            provider=LlmProviderNames.OPENAI,
+            api_key=None,
+            api_base=None,
+            api_version=None,
+            custom_config=None,
+            is_public=is_public,
+            model_configurations=[
+                ModelConfigurationUpsertRequest(
+                    name=default_model_name,
+                    is_visible=True,
+                )
+            ],
+        ),
+        db_session=db_session,
     )
-    db_session.add(provider)
-    db_session.flush()
+    if is_default:
+        update_default_provider(_provider.id, default_model_name, db_session)
+
+    provider = db_session.get(LLMProviderModel, _provider.id)
+    if not provider:
+        raise ValueError(f"Provider {name} not found")
     return provider
 
 
@@ -270,24 +277,6 @@ def test_get_llm_for_persona_falls_back_when_access_denied(
             provider_name=restricted_provider.name,
         )
 
-        # Set up ModelConfiguration + LLMModelFlow so get_default_llm() can
-        # resolve the default provider when the fallback path is triggered.
-        default_model_config = ModelConfiguration(
-            llm_provider_id=default_provider.id,
-            name=default_provider.default_model_name,
-            is_visible=True,
-        )
-        db_session.add(default_model_config)
-        db_session.flush()
-        db_session.add(
-            LLMModelFlow(
-                model_configuration_id=default_model_config.id,
-                llm_model_flow_type=LLMModelFlowType.CHAT,
-                is_default=True,
-            )
-        )
-        db_session.flush()
-
         access_group = UserGroup(name="persona-group")
         db_session.add(access_group)
         db_session.flush()
@@ -321,13 +310,19 @@ def test_get_llm_for_persona_falls_back_when_access_denied(
             persona=persona,
             user=admin_model,
         )
-        assert allowed_llm.config.model_name == restricted_provider.default_model_name
+        assert (
+            allowed_llm.config.model_name
+            == restricted_provider.model_configurations[0].name
+        )
 
         fallback_llm = get_llm_for_persona(
             persona=persona,
             user=basic_model,
         )
-        assert fallback_llm.config.model_name == default_provider.default_model_name
+        assert (
+            fallback_llm.config.model_name
+            == default_provider.model_configurations[0].name
+        )
 
 
 def test_list_llm_provider_basics_excludes_non_public_unrestricted(
@@ -346,6 +341,7 @@ def test_list_llm_provider_basics_excludes_non_public_unrestricted(
         name="public-provider",
         is_public=True,
         set_as_default=True,
+        default_model_name="gpt-4o",
         user_performing_action=admin_user,
     )
 
@@ -365,7 +361,7 @@ def test_list_llm_provider_basics_excludes_non_public_unrestricted(
         headers=basic_user.headers,
     )
     assert response.status_code == 200
-    providers = response.json()
+    providers = response.json()["providers"]
     provider_names = [p["name"] for p in providers]
 
     # Public provider should be visible
@@ -380,7 +376,7 @@ def test_list_llm_provider_basics_excludes_non_public_unrestricted(
         headers=admin_user.headers,
     )
     assert admin_response.status_code == 200
-    admin_providers = admin_response.json()
+    admin_providers = admin_response.json()["providers"]
     admin_provider_names = [p["name"] for p in admin_providers]
 
     assert public_provider.name in admin_provider_names
@@ -396,6 +392,7 @@ def test_provider_delete_clears_persona_references(reset: None) -> None:  # noqa
         name="default-provider",
         is_public=True,
         set_as_default=True,
+        default_model_name="gpt-4o",
         user_performing_action=admin_user,
     )
 

backend/tests/integration/tests/llm_provider/test_llm_provider_persona_access.py

@@ -107,7 +107,7 @@ def test_authorized_persona_access_returns_filtered_providers(
 
     # Should succeed
     assert response.status_code == 200
-    providers = response.json()
+    providers = response.json()["providers"]
 
     # Should include the restricted provider since basic_user can access the persona
     provider_names = [p["name"] for p in providers]
@@ -140,7 +140,7 @@ def test_persona_id_zero_applies_rbac(
 
     # Should succeed (persona_id=0 refers to default persona, which is public)
     assert response.status_code == 200
-    providers = response.json()
+    providers = response.json()["providers"]
 
     # Should NOT include the restricted provider since basic_user is not in group2
     provider_names = [p["name"] for p in providers]
@@ -182,7 +182,7 @@ def test_admin_can_query_any_persona(
 
     # Should succeed - admins can access any persona
     assert response.status_code == 200
-    providers = response.json()
+    providers = response.json()["providers"]
 
     # Should include the restricted provider
     provider_names = [p["name"] for p in providers]
@@ -223,7 +223,7 @@ def test_public_persona_accessible_to_all(
 
     # Should succeed
     assert response.status_code == 200
-    providers = response.json()
+    providers = response.json()["providers"]
 
     # Should return the public provider
     assert len(providers) > 0

backend/tests/unit/ee/onyx/server/settings/test_license_enforcement_settings.py

@@ -9,6 +9,19 @@ from redis.exceptions import RedisError
 from onyx.server.settings.models import ApplicationStatus
 from onyx.server.settings.models import Settings
 
+# Fields we assert on across all tests
+_ASSERT_FIELDS = {
+    "application_status",
+    "ee_features_enabled",
+    "seat_count",
+    "used_seats",
+}
+
+
+def _pick(settings: Settings) -> dict:
+    """Extract only the fields under test from a Settings object."""
+    return settings.model_dump(include=_ASSERT_FIELDS)
+
 
 @pytest.fixture
 def base_settings() -> Settings:
@@ -27,17 +40,17 @@ class TestApplyLicenseStatusToSettings:
     def test_enforcement_disabled_enables_ee_features(
         self, base_settings: Settings
     ) -> None:
-        """When LICENSE_ENFORCEMENT_ENABLED=False, EE features are enabled.
-
-        If we're running the EE apply function, EE code was loaded via
-        ENABLE_PAID_ENTERPRISE_EDITION_FEATURES, so features should be on.
-        """
+        """When LICENSE_ENFORCEMENT_ENABLED=False, EE features are enabled."""
         from ee.onyx.server.settings.api import apply_license_status_to_settings
 
         assert base_settings.ee_features_enabled is False
         result = apply_license_status_to_settings(base_settings)
-        assert result.application_status == ApplicationStatus.ACTIVE
-        assert result.ee_features_enabled is True
+        assert _pick(result) == {
+            "application_status": ApplicationStatus.ACTIVE,
+            "ee_features_enabled": True,
+            "seat_count": None,
+            "used_seats": None,
+        }
 
     @patch("ee.onyx.server.settings.api.LICENSE_ENFORCEMENT_ENABLED", True)
     @patch("ee.onyx.server.settings.api.MULTI_TENANT", True)
@@ -46,13 +59,60 @@ class TestApplyLicenseStatusToSettings:
         from ee.onyx.server.settings.api import apply_license_status_to_settings
 
         result = apply_license_status_to_settings(base_settings)
-        assert result.ee_features_enabled is True
+        assert _pick(result) == {
+            "application_status": ApplicationStatus.ACTIVE,
+            "ee_features_enabled": True,
+            "seat_count": None,
+            "used_seats": None,
+        }
 
     @pytest.mark.parametrize(
-        "license_status,expected_app_status,expected_ee_enabled",
+        "license_status,used_seats,seats,expected",
         [
-            (ApplicationStatus.GATED_ACCESS, ApplicationStatus.GATED_ACCESS, False),
-            (ApplicationStatus.ACTIVE, ApplicationStatus.ACTIVE, True),
+            (
+                ApplicationStatus.GATED_ACCESS,
+                3,
+                10,
+                {
+                    "application_status": ApplicationStatus.GATED_ACCESS,
+                    "ee_features_enabled": False,
+                    "seat_count": None,
+                    "used_seats": None,
+                },
+            ),
+            (
+                ApplicationStatus.ACTIVE,
+                3,
+                10,
+                {
+                    "application_status": ApplicationStatus.ACTIVE,
+                    "ee_features_enabled": True,
+                    "seat_count": None,
+                    "used_seats": None,
+                },
+            ),
+            (
+                ApplicationStatus.ACTIVE,
+                10,
+                10,
+                {
+                    "application_status": ApplicationStatus.ACTIVE,
+                    "ee_features_enabled": True,
+                    "seat_count": None,
+                    "used_seats": None,
+                },
+            ),
+            (
+                ApplicationStatus.GRACE_PERIOD,
+                3,
+                10,
+                {
+                    "application_status": ApplicationStatus.ACTIVE,
+                    "ee_features_enabled": True,
+                    "seat_count": None,
+                    "used_seats": None,
+                },
+            ),
         ],
     )
     @patch("ee.onyx.server.settings.api.LICENSE_ENFORCEMENT_ENABLED", True)
@@ -63,25 +123,80 @@ class TestApplyLicenseStatusToSettings:
         self,
         mock_get_metadata: MagicMock,
         mock_get_tenant: MagicMock,
-        license_status: ApplicationStatus | None,
-        expected_app_status: ApplicationStatus,
-        expected_ee_enabled: bool,
+        license_status: ApplicationStatus,
+        used_seats: int,
+        seats: int,
+        expected: dict,
         base_settings: Settings,
     ) -> None:
         """Self-hosted: license status controls both application_status and ee_features_enabled."""
         from ee.onyx.server.settings.api import apply_license_status_to_settings
 
         mock_get_tenant.return_value = "test_tenant"
-        if license_status is None:
-            mock_get_metadata.return_value = None
-        else:
-            mock_metadata = MagicMock()
-            mock_metadata.status = license_status
-            mock_get_metadata.return_value = mock_metadata
+        mock_metadata = MagicMock()
+        mock_metadata.status = license_status
+        mock_metadata.used_seats = used_seats
+        mock_metadata.seats = seats
+        mock_get_metadata.return_value = mock_metadata
 
         result = apply_license_status_to_settings(base_settings)
-        assert result.application_status == expected_app_status
-        assert result.ee_features_enabled is expected_ee_enabled
+        assert _pick(result) == expected
+
+    @patch("ee.onyx.server.settings.api.LICENSE_ENFORCEMENT_ENABLED", True)
+    @patch("ee.onyx.server.settings.api.MULTI_TENANT", False)
+    @patch("ee.onyx.server.settings.api.get_current_tenant_id")
+    @patch("ee.onyx.server.settings.api.get_cached_license_metadata")
+    def test_seat_limit_exceeded_sets_status_and_counts(
+        self,
+        mock_get_metadata: MagicMock,
+        mock_get_tenant: MagicMock,
+        base_settings: Settings,
+    ) -> None:
+        """Seat limit exceeded sets SEAT_LIMIT_EXCEEDED with counts, keeps EE enabled."""
+        from ee.onyx.server.settings.api import apply_license_status_to_settings
+
+        mock_get_tenant.return_value = "test_tenant"
+        mock_metadata = MagicMock()
+        mock_metadata.status = ApplicationStatus.ACTIVE
+        mock_metadata.used_seats = 15
+        mock_metadata.seats = 10
+        mock_get_metadata.return_value = mock_metadata
+
+        result = apply_license_status_to_settings(base_settings)
+        assert _pick(result) == {
+            "application_status": ApplicationStatus.SEAT_LIMIT_EXCEEDED,
+            "ee_features_enabled": True,
+            "seat_count": 10,
+            "used_seats": 15,
+        }
+
+    @patch("ee.onyx.server.settings.api.LICENSE_ENFORCEMENT_ENABLED", True)
+    @patch("ee.onyx.server.settings.api.MULTI_TENANT", False)
+    @patch("ee.onyx.server.settings.api.get_current_tenant_id")
+    @patch("ee.onyx.server.settings.api.get_cached_license_metadata")
+    def test_expired_license_takes_precedence_over_seat_limit(
+        self,
+        mock_get_metadata: MagicMock,
+        mock_get_tenant: MagicMock,
+        base_settings: Settings,
+    ) -> None:
+        """Expired license (GATED_ACCESS) takes precedence over seat limit exceeded."""
+        from ee.onyx.server.settings.api import apply_license_status_to_settings
+
+        mock_get_tenant.return_value = "test_tenant"
+        mock_metadata = MagicMock()
+        mock_metadata.status = ApplicationStatus.GATED_ACCESS
+        mock_metadata.used_seats = 15
+        mock_metadata.seats = 10
+        mock_get_metadata.return_value = mock_metadata
+
+        result = apply_license_status_to_settings(base_settings)
+        assert _pick(result) == {
+            "application_status": ApplicationStatus.GATED_ACCESS,
+            "ee_features_enabled": False,
+            "seat_count": None,
+            "used_seats": None,
+        }
 
     @patch("ee.onyx.server.settings.api.ENTERPRISE_EDITION_ENABLED", True)
     @patch("ee.onyx.server.settings.api.LICENSE_ENFORCEMENT_ENABLED", True)
@@ -105,8 +220,12 @@ class TestApplyLicenseStatusToSettings:
         mock_get_metadata.return_value = None
 
         result = apply_license_status_to_settings(base_settings)
-        assert result.application_status == ApplicationStatus.GATED_ACCESS
-        assert result.ee_features_enabled is False
+        assert _pick(result) == {
+            "application_status": ApplicationStatus.GATED_ACCESS,
+            "ee_features_enabled": False,
+            "seat_count": None,
+            "used_seats": None,
+        }
 
     @patch("ee.onyx.server.settings.api.ENTERPRISE_EDITION_ENABLED", False)
     @patch("ee.onyx.server.settings.api.LICENSE_ENFORCEMENT_ENABLED", True)
@@ -130,8 +249,12 @@ class TestApplyLicenseStatusToSettings:
         mock_get_metadata.return_value = None
 
         result = apply_license_status_to_settings(base_settings)
-        assert result.application_status == ApplicationStatus.ACTIVE
-        assert result.ee_features_enabled is False
+        assert _pick(result) == {
+            "application_status": ApplicationStatus.ACTIVE,
+            "ee_features_enabled": False,
+            "seat_count": None,
+            "used_seats": None,
+        }
 
     @patch("ee.onyx.server.settings.api.LICENSE_ENFORCEMENT_ENABLED", True)
     @patch("ee.onyx.server.settings.api.MULTI_TENANT", False)
@@ -150,8 +273,12 @@ class TestApplyLicenseStatusToSettings:
         mock_get_metadata.side_effect = RedisError("Connection failed")
 
         result = apply_license_status_to_settings(base_settings)
-        assert result.application_status == ApplicationStatus.ACTIVE
-        assert result.ee_features_enabled is False
+        assert _pick(result) == {
+            "application_status": ApplicationStatus.ACTIVE,
+            "ee_features_enabled": False,
+            "seat_count": None,
+            "used_seats": None,
+        }
 
 
 class TestSettingsDefaultEEDisabled:

backend/tests/unit/onyx/background/celery/tasks/test_user_file_impl_redis_locking.py

@@ -0,0 +1,291 @@
+"""Tests for the _impl functions' redis_locking parameter.
+
+Verifies that:
+- redis_locking=True acquires/releases Redis locks and clears queued keys
+- redis_locking=False skips all Redis operations entirely
+- Both paths execute the same business logic (DB lookup, status check)
+"""
+
+from unittest.mock import MagicMock
+from unittest.mock import patch
+from uuid import uuid4
+
+from onyx.background.celery.tasks.user_file_processing.tasks import (
+    _delete_user_file_impl,
+)
+from onyx.background.celery.tasks.user_file_processing.tasks import (
+    _process_user_file_impl,
+)
+from onyx.background.celery.tasks.user_file_processing.tasks import (
+    _project_sync_user_file_impl,
+)
+
+TASKS_MODULE = "onyx.background.celery.tasks.user_file_processing.tasks"
+
+
+def _mock_session_returning_none() -> MagicMock:
+    """Return a mock session whose .get() returns None (file not found)."""
+    session = MagicMock()
+    session.get.return_value = None
+    session.execute.return_value.scalar_one_or_none.return_value = None
+    return session
+
+
+# ------------------------------------------------------------------
+# _process_user_file_impl
+# ------------------------------------------------------------------
+
+
+class TestProcessUserFileImpl:
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    @patch(f"{TASKS_MODULE}.get_redis_client")
+    def test_redis_locking_true_acquires_and_releases_lock(
+        self,
+        mock_get_redis: MagicMock,
+        mock_get_session: MagicMock,
+    ) -> None:
+        redis_client = MagicMock()
+        lock = MagicMock()
+        lock.acquire.return_value = True
+        lock.owned.return_value = True
+        redis_client.lock.return_value = lock
+        mock_get_redis.return_value = redis_client
+
+        session = _mock_session_returning_none()
+        mock_get_session.return_value.__enter__.return_value = session
+
+        user_file_id = str(uuid4())
+        _process_user_file_impl(
+            user_file_id=user_file_id,
+            tenant_id="test-tenant",
+            redis_locking=True,
+        )
+
+        mock_get_redis.assert_called_once_with(tenant_id="test-tenant")
+        redis_client.delete.assert_called_once()
+        lock.acquire.assert_called_once_with(blocking=False)
+        lock.release.assert_called_once()
+
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    @patch(f"{TASKS_MODULE}.get_redis_client")
+    def test_redis_locking_true_skips_when_lock_held(
+        self,
+        mock_get_redis: MagicMock,
+        mock_get_session: MagicMock,
+    ) -> None:
+        redis_client = MagicMock()
+        lock = MagicMock()
+        lock.acquire.return_value = False
+        redis_client.lock.return_value = lock
+        mock_get_redis.return_value = redis_client
+
+        _process_user_file_impl(
+            user_file_id=str(uuid4()),
+            tenant_id="test-tenant",
+            redis_locking=True,
+        )
+
+        lock.acquire.assert_called_once()
+        mock_get_session.assert_not_called()
+
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    @patch(f"{TASKS_MODULE}.get_redis_client")
+    def test_redis_locking_false_skips_redis_entirely(
+        self,
+        mock_get_redis: MagicMock,
+        mock_get_session: MagicMock,
+    ) -> None:
+        session = _mock_session_returning_none()
+        mock_get_session.return_value.__enter__.return_value = session
+
+        _process_user_file_impl(
+            user_file_id=str(uuid4()),
+            tenant_id="test-tenant",
+            redis_locking=False,
+        )
+
+        mock_get_redis.assert_not_called()
+        mock_get_session.assert_called_once()
+
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    @patch(f"{TASKS_MODULE}.get_redis_client")
+    def test_both_paths_call_db_get(
+        self,
+        mock_get_redis: MagicMock,
+        mock_get_session: MagicMock,
+    ) -> None:
+        """Both redis_locking=True and False should call db_session.get(UserFile, ...)."""
+        redis_client = MagicMock()
+        lock = MagicMock()
+        lock.acquire.return_value = True
+        lock.owned.return_value = True
+        redis_client.lock.return_value = lock
+        mock_get_redis.return_value = redis_client
+
+        session = _mock_session_returning_none()
+        mock_get_session.return_value.__enter__.return_value = session
+
+        uid = str(uuid4())
+
+        _process_user_file_impl(user_file_id=uid, tenant_id="t", redis_locking=True)
+        call_count_true = session.get.call_count
+
+        session.reset_mock()
+        mock_get_session.reset_mock()
+        mock_get_session.return_value.__enter__.return_value = session
+
+        _process_user_file_impl(user_file_id=uid, tenant_id="t", redis_locking=False)
+        call_count_false = session.get.call_count
+
+        assert call_count_true == call_count_false == 1
+
+
+# ------------------------------------------------------------------
+# _delete_user_file_impl
+# ------------------------------------------------------------------
+
+
+class TestDeleteUserFileImpl:
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    @patch(f"{TASKS_MODULE}.get_redis_client")
+    def test_redis_locking_true_acquires_and_releases_lock(
+        self,
+        mock_get_redis: MagicMock,
+        mock_get_session: MagicMock,
+    ) -> None:
+        redis_client = MagicMock()
+        lock = MagicMock()
+        lock.acquire.return_value = True
+        lock.owned.return_value = True
+        redis_client.lock.return_value = lock
+        mock_get_redis.return_value = redis_client
+
+        session = _mock_session_returning_none()
+        mock_get_session.return_value.__enter__.return_value = session
+
+        _delete_user_file_impl(
+            user_file_id=str(uuid4()),
+            tenant_id="test-tenant",
+            redis_locking=True,
+        )
+
+        mock_get_redis.assert_called_once()
+        lock.acquire.assert_called_once_with(blocking=False)
+        lock.release.assert_called_once()
+
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    @patch(f"{TASKS_MODULE}.get_redis_client")
+    def test_redis_locking_true_skips_when_lock_held(
+        self,
+        mock_get_redis: MagicMock,
+        mock_get_session: MagicMock,
+    ) -> None:
+        redis_client = MagicMock()
+        lock = MagicMock()
+        lock.acquire.return_value = False
+        redis_client.lock.return_value = lock
+        mock_get_redis.return_value = redis_client
+
+        _delete_user_file_impl(
+            user_file_id=str(uuid4()),
+            tenant_id="test-tenant",
+            redis_locking=True,
+        )
+
+        lock.acquire.assert_called_once()
+        mock_get_session.assert_not_called()
+
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    @patch(f"{TASKS_MODULE}.get_redis_client")
+    def test_redis_locking_false_skips_redis_entirely(
+        self,
+        mock_get_redis: MagicMock,
+        mock_get_session: MagicMock,
+    ) -> None:
+        session = _mock_session_returning_none()
+        mock_get_session.return_value.__enter__.return_value = session
+
+        _delete_user_file_impl(
+            user_file_id=str(uuid4()),
+            tenant_id="test-tenant",
+            redis_locking=False,
+        )
+
+        mock_get_redis.assert_not_called()
+        mock_get_session.assert_called_once()
+
+
+# ------------------------------------------------------------------
+# _project_sync_user_file_impl
+# ------------------------------------------------------------------
+
+
+class TestProjectSyncUserFileImpl:
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    @patch(f"{TASKS_MODULE}.get_redis_client")
+    def test_redis_locking_true_acquires_and_releases_lock(
+        self,
+        mock_get_redis: MagicMock,
+        mock_get_session: MagicMock,
+    ) -> None:
+        redis_client = MagicMock()
+        lock = MagicMock()
+        lock.acquire.return_value = True
+        lock.owned.return_value = True
+        redis_client.lock.return_value = lock
+        mock_get_redis.return_value = redis_client
+
+        session = _mock_session_returning_none()
+        mock_get_session.return_value.__enter__.return_value = session
+
+        _project_sync_user_file_impl(
+            user_file_id=str(uuid4()),
+            tenant_id="test-tenant",
+            redis_locking=True,
+        )
+
+        mock_get_redis.assert_called_once()
+        redis_client.delete.assert_called_once()
+        lock.acquire.assert_called_once_with(blocking=False)
+        lock.release.assert_called_once()
+
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    @patch(f"{TASKS_MODULE}.get_redis_client")
+    def test_redis_locking_true_skips_when_lock_held(
+        self,
+        mock_get_redis: MagicMock,
+        mock_get_session: MagicMock,
+    ) -> None:
+        redis_client = MagicMock()
+        lock = MagicMock()
+        lock.acquire.return_value = False
+        redis_client.lock.return_value = lock
+        mock_get_redis.return_value = redis_client
+
+        _project_sync_user_file_impl(
+            user_file_id=str(uuid4()),
+            tenant_id="test-tenant",
+            redis_locking=True,
+        )
+
+        lock.acquire.assert_called_once()
+        mock_get_session.assert_not_called()
+
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    @patch(f"{TASKS_MODULE}.get_redis_client")
+    def test_redis_locking_false_skips_redis_entirely(
+        self,
+        mock_get_redis: MagicMock,
+        mock_get_session: MagicMock,
+    ) -> None:
+        session = _mock_session_returning_none()
+        mock_get_session.return_value.__enter__.return_value = session
+
+        _project_sync_user_file_impl(
+            user_file_id=str(uuid4()),
+            tenant_id="test-tenant",
+            redis_locking=False,
+        )
+
+        mock_get_redis.assert_not_called()
+        mock_get_session.assert_called_once()

backend/tests/unit/onyx/background/celery/tasks/test_user_file_processing_no_vectordb.py

@@ -0,0 +1,421 @@
+"""Tests for no-vector-DB user file processing paths.
+
+Verifies that when DISABLE_VECTOR_DB is True:
+- _process_user_file_impl calls _process_user_file_without_vector_db (not indexing)
+- _process_user_file_without_vector_db extracts text, counts tokens, stores plaintext,
+  sets status=COMPLETED and chunk_count=0
+- _delete_user_file_impl skips vector DB chunk deletion
+- _project_sync_user_file_impl skips vector DB metadata update
+"""
+
+from unittest.mock import MagicMock
+from unittest.mock import patch
+from uuid import uuid4
+
+from onyx.background.celery.tasks.user_file_processing.tasks import (
+    _delete_user_file_impl,
+)
+from onyx.background.celery.tasks.user_file_processing.tasks import (
+    _process_user_file_impl,
+)
+from onyx.background.celery.tasks.user_file_processing.tasks import (
+    _process_user_file_without_vector_db,
+)
+from onyx.background.celery.tasks.user_file_processing.tasks import (
+    _project_sync_user_file_impl,
+)
+from onyx.configs.constants import DocumentSource
+from onyx.connectors.models import Document
+from onyx.connectors.models import TextSection
+from onyx.db.enums import UserFileStatus
+
+TASKS_MODULE = "onyx.background.celery.tasks.user_file_processing.tasks"
+LLM_FACTORY_MODULE = "onyx.llm.factory"
+
+
+def _make_documents(texts: list[str]) -> list[Document]:
+    """Build a list of Document objects with the given section texts."""
+    return [
+        Document(
+            id=str(uuid4()),
+            source=DocumentSource.USER_FILE,
+            sections=[TextSection(text=t)],
+            semantic_identifier=f"test-doc-{i}",
+            metadata={},
+        )
+        for i, t in enumerate(texts)
+    ]
+
+
+def _make_user_file(
+    *,
+    status: UserFileStatus = UserFileStatus.PROCESSING,
+    file_id: str = "test-file-id",
+    name: str = "test.txt",
+) -> MagicMock:
+    """Return a MagicMock mimicking a UserFile ORM instance."""
+    uf = MagicMock()
+    uf.id = uuid4()
+    uf.file_id = file_id
+    uf.name = name
+    uf.status = status
+    uf.token_count = None
+    uf.chunk_count = None
+    uf.last_project_sync_at = None
+    uf.projects = []
+    uf.assistants = []
+    uf.needs_project_sync = True
+    uf.needs_persona_sync = True
+    return uf
+
+
+# ------------------------------------------------------------------
+# _process_user_file_without_vector_db — direct tests
+# ------------------------------------------------------------------
+
+
+class TestProcessUserFileWithoutVectorDb:
+    @patch(f"{TASKS_MODULE}.store_user_file_plaintext")
+    @patch(f"{LLM_FACTORY_MODULE}.get_llm_tokenizer_encode_func")
+    @patch(f"{LLM_FACTORY_MODULE}.get_default_llm")
+    def test_extracts_and_combines_text(
+        self,
+        mock_get_llm: MagicMock,  # noqa: ARG002
+        mock_get_encode: MagicMock,
+        mock_store_plaintext: MagicMock,
+    ) -> None:
+        mock_encode = MagicMock(return_value=[1, 2, 3, 4, 5])
+        mock_get_encode.return_value = mock_encode
+
+        uf = _make_user_file()
+        docs = _make_documents(["hello world", "foo bar"])
+        db_session = MagicMock()
+
+        _process_user_file_without_vector_db(uf, docs, db_session)
+
+        stored_text = mock_store_plaintext.call_args.kwargs["plaintext_content"]
+        assert "hello world" in stored_text
+        assert "foo bar" in stored_text
+
+    @patch(f"{TASKS_MODULE}.store_user_file_plaintext")
+    @patch(f"{LLM_FACTORY_MODULE}.get_llm_tokenizer_encode_func")
+    @patch(f"{LLM_FACTORY_MODULE}.get_default_llm")
+    def test_computes_token_count(
+        self,
+        mock_get_llm: MagicMock,  # noqa: ARG002
+        mock_get_encode: MagicMock,
+        mock_store_plaintext: MagicMock,  # noqa: ARG002
+    ) -> None:
+        mock_encode = MagicMock(return_value=list(range(42)))
+        mock_get_encode.return_value = mock_encode
+
+        uf = _make_user_file()
+        docs = _make_documents(["some text content"])
+        db_session = MagicMock()
+
+        _process_user_file_without_vector_db(uf, docs, db_session)
+
+        assert uf.token_count == 42
+
+    @patch(f"{TASKS_MODULE}.store_user_file_plaintext")
+    @patch(f"{LLM_FACTORY_MODULE}.get_llm_tokenizer_encode_func")
+    @patch(f"{LLM_FACTORY_MODULE}.get_default_llm")
+    def test_token_count_falls_back_to_none_on_error(
+        self,
+        mock_get_llm: MagicMock,
+        mock_get_encode: MagicMock,  # noqa: ARG002
+        mock_store_plaintext: MagicMock,  # noqa: ARG002
+    ) -> None:
+        mock_get_llm.side_effect = RuntimeError("No LLM configured")
+
+        uf = _make_user_file()
+        docs = _make_documents(["text"])
+        db_session = MagicMock()
+
+        _process_user_file_without_vector_db(uf, docs, db_session)
+
+        assert uf.token_count is None
+
+    @patch(f"{TASKS_MODULE}.store_user_file_plaintext")
+    @patch(f"{LLM_FACTORY_MODULE}.get_llm_tokenizer_encode_func")
+    @patch(f"{LLM_FACTORY_MODULE}.get_default_llm")
+    def test_stores_plaintext(
+        self,
+        mock_get_llm: MagicMock,  # noqa: ARG002
+        mock_get_encode: MagicMock,
+        mock_store_plaintext: MagicMock,
+    ) -> None:
+        mock_get_encode.return_value = MagicMock(return_value=[1])
+
+        uf = _make_user_file()
+        docs = _make_documents(["content to store"])
+        db_session = MagicMock()
+
+        _process_user_file_without_vector_db(uf, docs, db_session)
+
+        mock_store_plaintext.assert_called_once_with(
+            user_file_id=uf.id,
+            plaintext_content="content to store",
+        )
+
+    @patch(f"{TASKS_MODULE}.store_user_file_plaintext")
+    @patch(f"{LLM_FACTORY_MODULE}.get_llm_tokenizer_encode_func")
+    @patch(f"{LLM_FACTORY_MODULE}.get_default_llm")
+    def test_sets_completed_status_and_zero_chunk_count(
+        self,
+        mock_get_llm: MagicMock,  # noqa: ARG002
+        mock_get_encode: MagicMock,
+        mock_store_plaintext: MagicMock,  # noqa: ARG002
+    ) -> None:
+        mock_get_encode.return_value = MagicMock(return_value=[1])
+
+        uf = _make_user_file()
+        docs = _make_documents(["text"])
+        db_session = MagicMock()
+
+        _process_user_file_without_vector_db(uf, docs, db_session)
+
+        assert uf.status == UserFileStatus.COMPLETED
+        assert uf.chunk_count == 0
+        assert uf.last_project_sync_at is not None
+        db_session.add.assert_called_once_with(uf)
+        db_session.commit.assert_called_once()
+
+    @patch(f"{TASKS_MODULE}.store_user_file_plaintext")
+    @patch(f"{LLM_FACTORY_MODULE}.get_llm_tokenizer_encode_func")
+    @patch(f"{LLM_FACTORY_MODULE}.get_default_llm")
+    def test_preserves_deleting_status(
+        self,
+        mock_get_llm: MagicMock,  # noqa: ARG002
+        mock_get_encode: MagicMock,
+        mock_store_plaintext: MagicMock,  # noqa: ARG002
+    ) -> None:
+        mock_get_encode.return_value = MagicMock(return_value=[1])
+
+        uf = _make_user_file(status=UserFileStatus.DELETING)
+        docs = _make_documents(["text"])
+        db_session = MagicMock()
+
+        _process_user_file_without_vector_db(uf, docs, db_session)
+
+        assert uf.status == UserFileStatus.DELETING
+        assert uf.chunk_count == 0
+
+
+# ------------------------------------------------------------------
+# _process_user_file_impl — branching on DISABLE_VECTOR_DB
+# ------------------------------------------------------------------
+
+
+class TestProcessImplBranching:
+    @patch(f"{TASKS_MODULE}._process_user_file_without_vector_db")
+    @patch(f"{TASKS_MODULE}._process_user_file_with_indexing")
+    @patch(f"{TASKS_MODULE}.DISABLE_VECTOR_DB", True)
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    def test_calls_without_vector_db_when_disabled(
+        self,
+        mock_get_session: MagicMock,
+        mock_with_indexing: MagicMock,
+        mock_without_vdb: MagicMock,
+    ) -> None:
+        uf = _make_user_file()
+        session = MagicMock()
+        session.get.return_value = uf
+        mock_get_session.return_value.__enter__.return_value = session
+
+        connector_mock = MagicMock()
+        connector_mock.load_from_state.return_value = [_make_documents(["hello"])]
+
+        with patch(f"{TASKS_MODULE}.LocalFileConnector", return_value=connector_mock):
+            _process_user_file_impl(
+                user_file_id=str(uf.id),
+                tenant_id="test-tenant",
+                redis_locking=False,
+            )
+
+        mock_without_vdb.assert_called_once()
+        mock_with_indexing.assert_not_called()
+
+    @patch(f"{TASKS_MODULE}._process_user_file_without_vector_db")
+    @patch(f"{TASKS_MODULE}._process_user_file_with_indexing")
+    @patch(f"{TASKS_MODULE}.DISABLE_VECTOR_DB", False)
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    def test_calls_with_indexing_when_vector_db_enabled(
+        self,
+        mock_get_session: MagicMock,
+        mock_with_indexing: MagicMock,
+        mock_without_vdb: MagicMock,
+    ) -> None:
+        uf = _make_user_file()
+        session = MagicMock()
+        session.get.return_value = uf
+        mock_get_session.return_value.__enter__.return_value = session
+
+        connector_mock = MagicMock()
+        connector_mock.load_from_state.return_value = [_make_documents(["hello"])]
+
+        with patch(f"{TASKS_MODULE}.LocalFileConnector", return_value=connector_mock):
+            _process_user_file_impl(
+                user_file_id=str(uf.id),
+                tenant_id="test-tenant",
+                redis_locking=False,
+            )
+
+        mock_with_indexing.assert_called_once()
+        mock_without_vdb.assert_not_called()
+
+    @patch(f"{TASKS_MODULE}.run_indexing_pipeline")
+    @patch(f"{TASKS_MODULE}.store_user_file_plaintext")
+    @patch(f"{TASKS_MODULE}.DISABLE_VECTOR_DB", True)
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    def test_indexing_pipeline_not_called_when_disabled(
+        self,
+        mock_get_session: MagicMock,
+        mock_store_plaintext: MagicMock,  # noqa: ARG002
+        mock_run_pipeline: MagicMock,
+    ) -> None:
+        """End-to-end: verify run_indexing_pipeline is never invoked."""
+        uf = _make_user_file()
+        session = MagicMock()
+        session.get.return_value = uf
+        mock_get_session.return_value.__enter__.return_value = session
+
+        connector_mock = MagicMock()
+        connector_mock.load_from_state.return_value = [_make_documents(["content"])]
+
+        with (
+            patch(f"{TASKS_MODULE}.LocalFileConnector", return_value=connector_mock),
+            patch(f"{LLM_FACTORY_MODULE}.get_default_llm"),
+            patch(
+                f"{LLM_FACTORY_MODULE}.get_llm_tokenizer_encode_func",
+                return_value=MagicMock(return_value=[1, 2, 3]),
+            ),
+        ):
+            _process_user_file_impl(
+                user_file_id=str(uf.id),
+                tenant_id="test-tenant",
+                redis_locking=False,
+            )
+
+        mock_run_pipeline.assert_not_called()
+
+
+# ------------------------------------------------------------------
+# _delete_user_file_impl — vector DB skip
+# ------------------------------------------------------------------
+
+
+class TestDeleteImplNoVectorDb:
+    @patch(f"{TASKS_MODULE}.DISABLE_VECTOR_DB", True)
+    @patch(f"{TASKS_MODULE}.get_default_file_store")
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    def test_skips_vector_db_deletion(
+        self,
+        mock_get_session: MagicMock,
+        mock_get_file_store: MagicMock,
+    ) -> None:
+        uf = _make_user_file(status=UserFileStatus.DELETING)
+        session = MagicMock()
+        session.get.return_value = uf
+        mock_get_session.return_value.__enter__.return_value = session
+        mock_get_file_store.return_value = MagicMock()
+
+        with (
+            patch(f"{TASKS_MODULE}.get_all_document_indices") as mock_get_indices,
+            patch(f"{TASKS_MODULE}.get_active_search_settings") as mock_get_ss,
+            patch(f"{TASKS_MODULE}.httpx_init_vespa_pool") as mock_vespa_pool,
+        ):
+            _delete_user_file_impl(
+                user_file_id=str(uf.id),
+                tenant_id="test-tenant",
+                redis_locking=False,
+            )
+
+            mock_get_indices.assert_not_called()
+            mock_get_ss.assert_not_called()
+            mock_vespa_pool.assert_not_called()
+
+        session.delete.assert_called_once_with(uf)
+        session.commit.assert_called_once()
+
+    @patch(f"{TASKS_MODULE}.DISABLE_VECTOR_DB", True)
+    @patch(f"{TASKS_MODULE}.get_default_file_store")
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    def test_still_deletes_file_store_and_db_record(
+        self,
+        mock_get_session: MagicMock,
+        mock_get_file_store: MagicMock,
+    ) -> None:
+        uf = _make_user_file(status=UserFileStatus.DELETING)
+        session = MagicMock()
+        session.get.return_value = uf
+        mock_get_session.return_value.__enter__.return_value = session
+
+        file_store = MagicMock()
+        mock_get_file_store.return_value = file_store
+
+        _delete_user_file_impl(
+            user_file_id=str(uf.id),
+            tenant_id="test-tenant",
+            redis_locking=False,
+        )
+
+        assert file_store.delete_file.call_count == 2
+        session.delete.assert_called_once_with(uf)
+        session.commit.assert_called_once()
+
+
+# ------------------------------------------------------------------
+# _project_sync_user_file_impl — vector DB skip
+# ------------------------------------------------------------------
+
+
+class TestProjectSyncImplNoVectorDb:
+    @patch(f"{TASKS_MODULE}.DISABLE_VECTOR_DB", True)
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    def test_skips_vector_db_update(
+        self,
+        mock_get_session: MagicMock,
+    ) -> None:
+        uf = _make_user_file(status=UserFileStatus.COMPLETED)
+        session = MagicMock()
+        session.execute.return_value.scalar_one_or_none.return_value = uf
+        mock_get_session.return_value.__enter__.return_value = session
+
+        with (
+            patch(f"{TASKS_MODULE}.get_all_document_indices") as mock_get_indices,
+            patch(f"{TASKS_MODULE}.get_active_search_settings") as mock_get_ss,
+            patch(f"{TASKS_MODULE}.httpx_init_vespa_pool") as mock_vespa_pool,
+        ):
+            _project_sync_user_file_impl(
+                user_file_id=str(uf.id),
+                tenant_id="test-tenant",
+                redis_locking=False,
+            )
+
+            mock_get_indices.assert_not_called()
+            mock_get_ss.assert_not_called()
+            mock_vespa_pool.assert_not_called()
+
+    @patch(f"{TASKS_MODULE}.DISABLE_VECTOR_DB", True)
+    @patch(f"{TASKS_MODULE}.get_session_with_current_tenant")
+    def test_still_clears_sync_flags(
+        self,
+        mock_get_session: MagicMock,
+    ) -> None:
+        uf = _make_user_file(status=UserFileStatus.COMPLETED)
+        session = MagicMock()
+        session.execute.return_value.scalar_one_or_none.return_value = uf
+        mock_get_session.return_value.__enter__.return_value = session
+
+        _project_sync_user_file_impl(
+            user_file_id=str(uf.id),
+            tenant_id="test-tenant",
+            redis_locking=False,
+        )
+
+        assert uf.needs_project_sync is False
+        assert uf.needs_persona_sync is False
+        assert uf.last_project_sync_at is not None
+        session.add.assert_called_once_with(uf)
+        session.commit.assert_called_once()

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

@@ -44,7 +44,6 @@ def _build_provider_view(
         id=1,
         name="test-provider",
         provider=provider,
-        default_model_name="test-model",
         model_configurations=[
             ModelConfigurationView(
                 name="test-model",
@@ -62,7 +61,6 @@ def _build_provider_view(
         groups=[],
         personas=[],
         deployment_name=None,
-        default_vision_model=None,
     )
 
 

backend/tests/unit/onyx/test_startup_validation.py

@@ -0,0 +1,52 @@
+"""Tests for startup validation in no-vector-DB mode.
+
+Verifies that DISABLE_VECTOR_DB raises RuntimeError when combined with
+incompatible settings (MULTI_TENANT, ENABLE_CRAFT).
+"""
+
+from unittest.mock import patch
+
+import pytest
+
+
+class TestValidateNoVectorDbSettings:
+    @patch("onyx.main.DISABLE_VECTOR_DB", False)
+    def test_no_error_when_vector_db_enabled(self) -> None:
+        from onyx.main import validate_no_vector_db_settings
+
+        validate_no_vector_db_settings()
+
+    @patch("onyx.main.DISABLE_VECTOR_DB", True)
+    @patch("onyx.main.MULTI_TENANT", False)
+    @patch("onyx.server.features.build.configs.ENABLE_CRAFT", False)
+    def test_no_error_when_no_conflicts(self) -> None:
+        from onyx.main import validate_no_vector_db_settings
+
+        validate_no_vector_db_settings()
+
+    @patch("onyx.main.DISABLE_VECTOR_DB", True)
+    @patch("onyx.main.MULTI_TENANT", True)
+    def test_raises_on_multi_tenant(self) -> None:
+        from onyx.main import validate_no_vector_db_settings
+
+        with pytest.raises(RuntimeError, match="MULTI_TENANT"):
+            validate_no_vector_db_settings()
+
+    @patch("onyx.main.DISABLE_VECTOR_DB", True)
+    @patch("onyx.main.MULTI_TENANT", False)
+    @patch("onyx.server.features.build.configs.ENABLE_CRAFT", True)
+    def test_raises_on_enable_craft(self) -> None:
+        from onyx.main import validate_no_vector_db_settings
+
+        with pytest.raises(RuntimeError, match="ENABLE_CRAFT"):
+            validate_no_vector_db_settings()
+
+    @patch("onyx.main.DISABLE_VECTOR_DB", True)
+    @patch("onyx.main.MULTI_TENANT", True)
+    @patch("onyx.server.features.build.configs.ENABLE_CRAFT", True)
+    def test_multi_tenant_checked_before_craft(self) -> None:
+        """MULTI_TENANT is checked first, so it should be the error raised."""
+        from onyx.main import validate_no_vector_db_settings
+
+        with pytest.raises(RuntimeError, match="MULTI_TENANT"):
+            validate_no_vector_db_settings()

deployment/docker_compose/docker-compose.multitenant-dev.yml

@@ -468,7 +468,7 @@ services:
       - minio_data:/data
     command: server /data --console-address ":9001"
     healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
+      test: ["CMD", "mc", "ready", "local"]
       interval: 30s
       timeout: 20s
       retries: 3

deployment/docker_compose/docker-compose.prod-cloud.yml

@@ -293,7 +293,7 @@ services:
       - minio_data:/data
     command: server /data --console-address ":9001"
     healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
+      test: ["CMD", "mc", "ready", "local"]
       interval: 30s
       timeout: 20s
       retries: 3

deployment/docker_compose/docker-compose.prod-no-letsencrypt.yml

@@ -298,7 +298,7 @@ services:
       - minio_data:/data
     command: server /data --console-address ":9001"
     healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
+      test: ["CMD", "mc", "ready", "local"]
       interval: 30s
       timeout: 20s
       retries: 3

deployment/docker_compose/docker-compose.prod.yml

@@ -335,7 +335,7 @@ services:
       - minio_data:/data
     command: server /data --console-address ":9001"
     healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
+      test: ["CMD", "mc", "ready", "local"]
       interval: 30s
       timeout: 20s
       retries: 3

deployment/docker_compose/docker-compose.search-testing.yml

@@ -232,7 +232,7 @@ services:
       - minio_data:/data
     command: server /data --console-address ":9001"
     healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
+      test: ["CMD", "mc", "ready", "local"]
       interval: 30s
       timeout: 20s
       retries: 3

deployment/docker_compose/docker-compose.yml

@@ -520,7 +520,7 @@ services:
       - minio_data:/data
     command: server /data --console-address ":9001"
     healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
+      test: ["CMD", "mc", "ready", "local"]
       interval: 30s
       timeout: 20s
       retries: 3

web/lib/opal/src/core/hoverable/styles.css

@@ -1,6 +1,6 @@
 /* Hoverable — item transitions */
 .hoverable-item {
-  transition: opacity 200ms ease-in-out;
+  transition: opacity 150ms ease-in-out;
 }
 
 .hoverable-item[data-hoverable-variant="opacity-on-hover"] {

web/lib/opal/src/layouts/ContentAction/components.tsx

@@ -58,7 +58,7 @@ function ContentAction({
 
   return (
     <div className="flex flex-row items-stretch w-full">
-      <div className={cn("flex-1 min-w-0", padding)}>
+      <div className={cn("flex-1 min-w-0 self-center", padding)}>
         <Content {...contentProps} />
       </div>
       {rightChildren && (

web/src/app/admin/billing/BillingDetailsView.tsx

@@ -2,7 +2,8 @@
 
 import { useState } from "react";
 import Link from "next/link";
-import { Section, LineItemLayout } from "@/layouts/general-layouts";
+import { Section } from "@/layouts/general-layouts";
+import { Content } from "@opal/layouts";
 import * as InputLayouts from "@/layouts/input-layouts";
 import Card from "@/refresh-components/cards/Card";
 import Button from "@/refresh-components/buttons/Button";
@@ -372,9 +373,11 @@ function SeatsCard({
           padding={1}
           height="auto"
         >
-          <LineItemLayout
+          <Content
             title="Update Seats"
             description="Add or remove seats to reflect your team size."
+            sizePreset="main-content"
+            variant="section"
           />
           <Button main secondary onClick={handleCancel} disabled={isSubmitting}>
             Cancel

web/src/app/admin/configuration/image-generation/ImageGenerationConnectionModal.tsx

@@ -2,7 +2,7 @@
 
 import { ModalCreationInterface } from "@/refresh-components/contexts/ModalContext";
 import { ImageProvider } from "@/app/admin/configuration/image-generation/constants";
-import { LLMProviderView } from "@/app/admin/configuration/llm/interfaces";
+import { LLMProviderView } from "@/interfaces/llm";
 import { ImageGenerationConfigView } from "@/lib/configuration/imageConfigurationService";
 import { getImageGenForm } from "./forms";
 

web/src/app/admin/configuration/image-generation/ImageGenerationContent.tsx

@@ -7,7 +7,7 @@ import { Select } from "@/refresh-components/cards";
 import { useCreateModal } from "@/refresh-components/contexts/ModalContext";
 import { toast } from "@/hooks/useToast";
 import { errorHandlingFetcher } from "@/lib/fetcher";
-import { LLMProviderView } from "@/app/admin/configuration/llm/interfaces";
+import { LLMProviderResponse, LLMProviderView } from "@/interfaces/llm";
 import {
   IMAGE_PROVIDER_GROUPS,
   ImageProvider,
@@ -23,13 +23,14 @@ import Message from "@/refresh-components/messages/Message";
 
 export default function ImageGenerationContent() {
   const {
-    data: llmProviders = [],
+    data: llmProviderResponse,
     error: llmError,
     mutate: refetchProviders,
-  } = useSWR<LLMProviderView[]>(
+  } = useSWR<LLMProviderResponse<LLMProviderView>>(
     "/api/admin/llm/provider?include_image_gen=true",
     errorHandlingFetcher
   );
+  const llmProviders = llmProviderResponse?.providers ?? [];
 
   const {
     data: configs = [],

web/src/app/admin/configuration/image-generation/forms/types.ts

@@ -1,6 +1,6 @@
 import { FormikProps } from "formik";
 import { ImageProvider } from "../constants";
-import { LLMProviderView } from "@/app/admin/configuration/llm/interfaces";
+import { LLMProviderView } from "@/interfaces/llm";
 import {
   ImageGenerationConfigView,
   ImageGenerationCredentials,

web/src/app/admin/configuration/llm/LLMConfiguration.tsx

@@ -1,84 +0,0 @@
-"use client";
-
-import { errorHandlingFetcher } from "@/lib/fetcher";
-import useSWR from "swr";
-import { Callout } from "@/components/ui/callout";
-import Text from "@/refresh-components/texts/Text";
-import Title from "@/components/ui/title";
-import { ThreeDotsLoader } from "@/components/Loading";
-import { LLMProviderView } from "./interfaces";
-import { LLM_PROVIDERS_ADMIN_URL } from "./constants";
-import { OpenAIForm } from "./forms/OpenAIForm";
-import { AnthropicForm } from "./forms/AnthropicForm";
-import { OllamaForm } from "./forms/OllamaForm";
-import { AzureForm } from "./forms/AzureForm";
-import { BedrockForm } from "./forms/BedrockForm";
-import { VertexAIForm } from "./forms/VertexAIForm";
-import { OpenRouterForm } from "./forms/OpenRouterForm";
-import { getFormForExistingProvider } from "./forms/getForm";
-import { CustomForm } from "./forms/CustomForm";
-
-export function LLMConfiguration() {
-  const { data: existingLlmProviders } = useSWR<LLMProviderView[]>(
-    LLM_PROVIDERS_ADMIN_URL,
-    errorHandlingFetcher
-  );
-
-  if (!existingLlmProviders) {
-    return <ThreeDotsLoader />;
-  }
-
-  const isFirstProvider = existingLlmProviders.length === 0;
-
-  return (
-    <>
-      <Title className="mb-2">Enabled LLM Providers</Title>
-
-      {existingLlmProviders.length > 0 ? (
-        <>
-          <Text as="p" className="mb-4">
-            If multiple LLM providers are enabled, the default provider will be
-            used for all &quot;Default&quot; Assistants. For user-created
-            Assistants, you can select the LLM provider/model that best fits the
-            use case!
-          </Text>
-          <div className="flex flex-col gap-y-4">
-            {[...existingLlmProviders]
-              .sort((a, b) => {
-                if (a.is_default_provider && !b.is_default_provider) return -1;
-                if (!a.is_default_provider && b.is_default_provider) return 1;
-                return 0;
-              })
-              .map((llmProvider) => (
-                <div key={llmProvider.id}>
-                  {getFormForExistingProvider(llmProvider)}
-                </div>
-              ))}
-          </div>
-        </>
-      ) : (
-        <Callout type="warning" title="No LLM providers configured yet">
-          Please set one up below in order to start using Onyx!
-        </Callout>
-      )}
-
-      <Title className="mb-2 mt-6">Add LLM Provider</Title>
-      <Text as="p" className="mb-4">
-        Add a new LLM provider by either selecting from one of the default
-        providers or by specifying your own custom LLM provider.
-      </Text>
-
-      <div className="flex flex-col gap-y-4">
-        <OpenAIForm shouldMarkAsDefault={isFirstProvider} />
-        <AnthropicForm shouldMarkAsDefault={isFirstProvider} />
-        <OllamaForm shouldMarkAsDefault={isFirstProvider} />
-        <AzureForm shouldMarkAsDefault={isFirstProvider} />
-        <BedrockForm shouldMarkAsDefault={isFirstProvider} />
-        <VertexAIForm shouldMarkAsDefault={isFirstProvider} />
-        <OpenRouterForm shouldMarkAsDefault={isFirstProvider} />
-
-        <CustomForm shouldMarkAsDefault={isFirstProvider} />
-      </div>
-    </>
-  );
-}

web/src/app/admin/configuration/llm/ModelConfigurationField.tsx

@@ -1,13 +1,14 @@
 "use client";
 
 import { ArrayHelpers, FieldArray, FormikProps, useField } from "formik";
-import { ModelConfiguration } from "./interfaces";
+import { ModelConfiguration } from "@/interfaces/llm";
 import { ManualErrorMessage, TextFormField } from "@/components/Field";
 import { useEffect, useState } from "react";
 import CreateButton from "@/refresh-components/buttons/CreateButton";
 import { Button } from "@opal/components";
 import { SvgX } from "@opal/icons";
 import Text from "@/refresh-components/texts/Text";
+
 function ModelConfigurationRow({
   name,
   index,

web/src/app/admin/configuration/llm/forms/getForm.tsx

@@ -1,44 +0,0 @@
-import { LLMProviderName, LLMProviderView } from "../interfaces";
-import { AnthropicForm } from "./AnthropicForm";
-import { OpenAIForm } from "./OpenAIForm";
-import { OllamaForm } from "./OllamaForm";
-import { AzureForm } from "./AzureForm";
-import { VertexAIForm } from "./VertexAIForm";
-import { OpenRouterForm } from "./OpenRouterForm";
-import { CustomForm } from "./CustomForm";
-import { BedrockForm } from "./BedrockForm";
-
-export function detectIfRealOpenAIProvider(provider: LLMProviderView) {
-  return (
-    provider.provider === LLMProviderName.OPENAI &&
-    provider.api_key &&
-    !provider.api_base &&
-    Object.keys(provider.custom_config || {}).length === 0
-  );
-}
-
-export const getFormForExistingProvider = (provider: LLMProviderView) => {
-  switch (provider.provider) {
-    case LLMProviderName.OPENAI:
-      // "openai" as a provider name can be used for litellm proxy / any OpenAI-compatible provider
-      if (detectIfRealOpenAIProvider(provider)) {
-        return <OpenAIForm existingLlmProvider={provider} />;
-      } else {
-        return <CustomForm existingLlmProvider={provider} />;
-      }
-    case LLMProviderName.ANTHROPIC:
-      return <AnthropicForm existingLlmProvider={provider} />;
-    case LLMProviderName.OLLAMA_CHAT:
-      return <OllamaForm existingLlmProvider={provider} />;
-    case LLMProviderName.AZURE:
-      return <AzureForm existingLlmProvider={provider} />;
-    case LLMProviderName.VERTEX_AI:
-      return <VertexAIForm existingLlmProvider={provider} />;
-    case LLMProviderName.BEDROCK:
-      return <BedrockForm existingLlmProvider={provider} />;
-    case LLMProviderName.OPENROUTER:
-      return <OpenRouterForm existingLlmProvider={provider} />;
-    default:
-      return <CustomForm existingLlmProvider={provider} />;
-  }
-};

web/src/app/admin/configuration/llm/page.tsx

@@ -1,14 +1,7 @@
 "use client";
 
-import { AdminPageTitle } from "@/components/admin/Title";
-import { LLMConfiguration } from "./LLMConfiguration";
-import { SvgCpu } from "@opal/icons";
-export default function Page() {
-  return (
-    <>
-      <AdminPageTitle title="LLM Setup" icon={SvgCpu} />
+import LLMConfigurationPage from "@/refresh-pages/admin/LLMConfigurationPage";
 
-      <LLMConfiguration />
-    </>
-  );
+export default function Page() {
+  return <LLMConfigurationPage />;
 }

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

@@ -24,7 +24,7 @@ import {
   BedrockFetchParams,
   OllamaFetchParams,
   OpenRouterFetchParams,
-} from "./interfaces";
+} from "@/interfaces/llm";
 import { SvgAws, SvgOpenrouter } from "@opal/icons";
 
 // Aggregator providers that host models from multiple vendors
@@ -106,8 +106,9 @@ export const getProviderIcon = (
   return CPUIcon;
 };
 
-export const isAnthropic = (provider: string, modelName: string) =>
-  provider === "anthropic" || modelName.toLowerCase().includes("claude");
+export const isAnthropic = (provider: string, modelName?: string) =>
+  provider === LLMProviderName.ANTHROPIC ||
+  !!modelName?.toLowerCase().includes("claude");
 
 /**
  * Fetches Bedrock models directly without any form state dependencies.
@@ -153,6 +154,7 @@ export const fetchBedrockModels = async (
       is_visible: false,
       max_input_tokens: modelData.max_input_tokens,
       supports_image_input: modelData.supports_image_input,
+      supports_reasoning: false,
     }));
 
     return { models };
@@ -205,6 +207,7 @@ export const fetchOllamaModels = async (
       is_visible: true,
       max_input_tokens: modelData.max_input_tokens,
       supports_image_input: modelData.supports_image_input,
+      supports_reasoning: false,
     }));
 
     return { models };
@@ -262,6 +265,7 @@ export const fetchOpenRouterModels = async (
       is_visible: true,
       max_input_tokens: modelData.max_input_tokens,
       supports_image_input: modelData.supports_image_input,
+      supports_reasoning: false,
     }));
 
     return { models };

web/src/app/admin/connectors/[connector]/pages/FieldRendering.tsx

@@ -12,6 +12,7 @@ import Tabs from "@/refresh-components/Tabs";
 import { useFormikContext } from "formik";
 import * as GeneralLayouts from "@/layouts/general-layouts";
 import * as InputLayouts from "@/layouts/input-layouts";
+import { Content } from "@opal/layouts";
 import CheckboxField from "@/refresh-components/form/LabeledCheckboxField";
 import InputTextAreaField from "@/refresh-components/form/InputTextAreaField";
 import Text from "@/refresh-components/texts/Text";
@@ -46,9 +47,11 @@ const TabsField: FC<TabsFieldProps> = ({
   return (
     <GeneralLayouts.Section gap={0.5} alignItems="start">
       {tabField.label && (
-        <InputLayouts.Title
+        <Content
           title={resolvedLabel ?? ""}
           description={resolvedDescription}
+          sizePreset="main-content"
+          variant="section"
         />
       )}
 

web/src/app/admin/discord-bot/[guild-id]/page.tsx

@@ -5,7 +5,8 @@ import { cn } from "@/lib/utils";
 import { ThreeDotsLoader } from "@/components/Loading";
 import { ErrorCallout } from "@/components/ErrorCallout";
 import { toast } from "@/hooks/useToast";
-import { Section, LineItemLayout } from "@/layouts/general-layouts";
+import { Section } from "@/layouts/general-layouts";
+import { ContentAction } from "@opal/layouts";
 import * as SettingsLayouts from "@/layouts/settings-layouts";
 import Text from "@/refresh-components/texts/Text";
 import Card from "@/refresh-components/cards/Card";
@@ -89,9 +90,11 @@ function GuildDetailContent({
       )}
 
       <Card variant={disabled ? "disabled" : "primary"}>
-        <LineItemLayout
+        <ContentAction
           title="Channel Configuration"
           description="Run !sync-channels in Discord to update the channel list."
+          sizePreset="main-content"
+          variant="section"
           rightChildren={
             isRegistered && !channelsLoading && !channelsError ? (
               <Section
@@ -339,9 +342,11 @@ export default function Page({ params }: Props) {
       <SettingsLayouts.Body>
         {/* Default Persona Selector */}
         <Card variant={!guild?.enabled ? "disabled" : "primary"}>
-          <LineItemLayout
+          <ContentAction
             title="Default Agent"
             description="The agent used by the bot in all channels unless overridden."
+            sizePreset="main-content"
+            variant="section"
             rightChildren={
               <InputSelect
                 value={guild?.default_persona_id?.toString() ?? "default"}

web/src/app/admin/document-index-migration/page.tsx

@@ -5,7 +5,7 @@ import useSWR from "swr";
 import { SvgArrowExchange } from "@opal/icons";
 import * as SettingsLayouts from "@/layouts/settings-layouts";
 import Card from "@/refresh-components/cards/Card";
-import { LineItemLayout } from "@/layouts/general-layouts";
+import { Content, ContentAction } from "@opal/layouts";
 import Text from "@/refresh-components/texts/Text";
 import InputSelect from "@/refresh-components/inputs/InputSelect";
 import Button from "@/refresh-components/buttons/Button";
@@ -72,9 +72,10 @@ function MigrationStatusSection() {
     <Card>
       <Text headingH3>Migration Status</Text>
 
-      <LineItemLayout
+      <ContentAction
         title="Started"
-        variant="secondary"
+        sizePreset="main-ui"
+        variant="section"
         rightChildren={
           <Text mainUiBody>
             {hasStarted ? formatTimestamp(data.created_at!) : "Not started"}
@@ -82,9 +83,10 @@ function MigrationStatusSection() {
         }
       />
 
-      <LineItemLayout
+      <ContentAction
         title="Chunks Migrated"
-        variant="secondary"
+        sizePreset="main-ui"
+        variant="section"
         rightChildren={
           <Text mainUiBody>
             {progressPercentage !== null
@@ -96,9 +98,10 @@ function MigrationStatusSection() {
         }
       />
 
-      <LineItemLayout
+      <ContentAction
         title="Completed"
-        variant="secondary"
+        sizePreset="main-ui"
+        variant="section"
         rightChildren={
           <Text mainUiBody>
             {hasCompleted
@@ -174,10 +177,11 @@ function RetrievalSourceSection() {
 
   return (
     <Card>
-      <LineItemLayout
+      <Content
         title="Retrieval Source"
         description="Controls which document index is used for retrieval."
-        variant="secondary"
+        sizePreset="main-ui"
+        variant="section"
       />
 
       <InputSelect

web/src/app/admin/embeddings/EmbeddingModelSelectionForm.tsx

@@ -25,7 +25,7 @@ import { ModelOption } from "@/components/embedding/ModelSelector";
 import {
   EMBEDDING_MODELS_ADMIN_URL,
   EMBEDDING_PROVIDERS_ADMIN_URL,
-} from "@/app/admin/configuration/llm/constants";
+} from "@/lib/llmConfig/constants";
 import { AdvancedSearchConfiguration } from "@/app/admin/embeddings/interfaces";
 import Button from "@/refresh-components/buttons/Button";
 

web/src/app/admin/embeddings/modals/ChangeCredentialsModal.tsx

@@ -14,7 +14,7 @@ import {
 import {
   EMBEDDING_PROVIDERS_ADMIN_URL,
   LLM_PROVIDERS_ADMIN_URL,
-} from "@/app/admin/configuration/llm/constants";
+} from "@/lib/llmConfig/constants";
 import { mutate } from "swr";
 import { testEmbedding } from "@/app/admin/embeddings/pages/utils";
 import { SvgSettings } from "@opal/icons";

web/src/app/admin/embeddings/modals/ProviderCreationModal.tsx

@@ -11,7 +11,7 @@ import {
   EmbeddingProvider,
   getFormattedProviderName,
 } from "@/components/embedding/interfaces";
-import { EMBEDDING_PROVIDERS_ADMIN_URL } from "@/app/admin/configuration/llm/constants";
+import { EMBEDDING_PROVIDERS_ADMIN_URL } from "@/lib/llmConfig/constants";
 import Modal from "@/refresh-components/Modal";
 import { SvgSettings } from "@opal/icons";
 export interface ProviderCreationModalProps {

web/src/app/admin/embeddings/pages/AdvancedEmbeddingFormPage.tsx

@@ -15,7 +15,7 @@ import {
 import NumberInput from "../../connectors/[connector]/pages/ConnectorInput/NumberInput";
 import { StringOrNumberOption } from "@/components/Dropdown";
 import useSWR from "swr";
-import { LLM_CONTEXTUAL_COST_ADMIN_URL } from "../../configuration/llm/constants";
+import { LLM_CONTEXTUAL_COST_ADMIN_URL } from "@/lib/llmConfig/constants";
 import { errorHandlingFetcher } from "@/lib/fetcher";
 import Button from "@/refresh-components/buttons/Button";
 import { NEXT_PUBLIC_CLOUD_ENABLED } from "@/lib/constants";

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

@@ -18,7 +18,7 @@ import SourceTag from "@/refresh-components/buttons/source-tag/SourceTag";
 import { citationsToSourceInfoArray } from "@/refresh-components/buttons/source-tag/sourceTagUtils";
 import CopyIconButton from "@/refresh-components/buttons/CopyIconButton";
 import LLMPopover from "@/refresh-components/popovers/LLMPopover";
-import { parseLlmDescriptor } from "@/lib/llm/utils";
+import { parseLlmDescriptor } from "@/lib/llmConfig/utils";
 import { LlmManager } from "@/lib/hooks";
 import { Message } from "@/app/app/interfaces";
 import { SvgThumbsDown, SvgThumbsUp } from "@opal/icons";

web/src/app/app/message/messageComponents/timeline/headers/CompletedHeader.tsx

@@ -6,7 +6,8 @@ import { Button } from "@opal/components";
 import Tag from "@/refresh-components/buttons/Tag";
 import Text from "@/refresh-components/texts/Text";
 import SimpleTooltip from "@/refresh-components/SimpleTooltip";
-import { Section, LineItemLayout } from "@/layouts/general-layouts";
+import { Section } from "@/layouts/general-layouts";
+import { ContentAction } from "@opal/layouts";
 import { formatDurationSeconds } from "@/lib/time";
 import { noProp } from "@/lib/utils";
 import MemoriesModal from "@/refresh-components/modals/MemoriesModal";
@@ -61,10 +62,12 @@ function MemoryTagWithTooltip({
                   {memoryText}
                 </Text>
               </div>
-              <LineItemLayout
-                variant="mini"
+              <ContentAction
                 icon={SvgAddLines}
                 title={operationLabel}
+                sizePreset="secondary"
+                variant="body"
+                prominence="muted"
                 rightChildren={
                   <Button
                     prominence="tertiary"

web/src/app/craft/components/BuildLLMPopover.tsx

@@ -11,7 +11,7 @@ import Text from "@/refresh-components/texts/Text";
 import Popover, { PopoverMenu } from "@/refresh-components/Popover";
 import Switch from "@/refresh-components/inputs/Switch";
 import LineItem from "@/refresh-components/buttons/LineItem";
-import { LLMProviderDescriptor } from "@/app/admin/configuration/llm/interfaces";
+import { LLMProviderDescriptor } from "@/interfaces/llm";
 import {
   BuildLlmSelection,
   BUILD_MODE_PROVIDERS,

web/src/app/craft/components/ShareButton.tsx

@@ -9,7 +9,8 @@ import type { SharingScope } from "@/app/craft/types/streamingTypes";
 import { cn } from "@/lib/utils";
 import Popover from "@/refresh-components/Popover";
 import Truncated from "@/refresh-components/texts/Truncated";
-import { Section, LineItemLayout } from "@/layouts/general-layouts";
+import { Section } from "@/layouts/general-layouts";
+import { ContentAction } from "@opal/layouts";
 
 interface ShareButtonProps {
   sessionId: string;
@@ -133,11 +134,12 @@ export default function ShareButton({
                       : "hover:bg-background-tint-02"
                   )}
                 >
-                  <LineItemLayout
+                  <ContentAction
                     title={opt.label}
                     description={opt.description}
-                    variant="tertiary"
-                    reducedPadding
+                    sizePreset="main-ui"
+                    variant="section"
+                    paddingVariant="sm"
                   />
                 </div>
               ))}

web/src/app/craft/hooks/useBuildLlmSelection.ts

@@ -1,5 +1,5 @@
 import { useMemo, useState, useCallback } from "react";
-import { LLMProviderDescriptor } from "@/app/admin/configuration/llm/interfaces";
+import { LLMProviderDescriptor } from "@/interfaces/llm";
 import {
   BuildLlmSelection,
   getBuildLlmSelection,

web/src/app/craft/hooks/useBuildSessionController.ts

@@ -7,7 +7,7 @@ import { usePreProvisionPolling } from "@/app/craft/hooks/usePreProvisionPolling
 import { CRAFT_SEARCH_PARAM_NAMES } from "@/app/craft/services/searchParams";
 import { CRAFT_PATH } from "@/app/craft/v1/constants";
 import { getBuildUserPersona } from "@/app/craft/onboarding/constants";
-import { useLLMProviders } from "@/lib/hooks/useLLMProviders";
+import { useLLMProviders } from "@/hooks/useLLMProviders";
 import { checkPreProvisionedSession } from "@/app/craft/services/apiServices";
 
 interface UseBuildSessionControllerProps {

web/src/app/craft/onboarding/components/BuildOnboardingModal.tsx

@@ -18,8 +18,8 @@ import {
   getBuildLlmSelection,
   getDefaultLlmSelection,
 } from "@/app/craft/onboarding/constants";
-import { LLMProviderDescriptor } from "@/app/admin/configuration/llm/interfaces";
-import { LLM_PROVIDERS_ADMIN_URL } from "@/app/admin/configuration/llm/constants";
+import { LLMProviderDescriptor } from "@/interfaces/llm";
+import { LLM_PROVIDERS_ADMIN_URL } from "@/lib/llmConfig/constants";
 import {
   buildInitialValues,
   testApiKeyHelper,

web/src/app/craft/onboarding/components/OnboardingLlmSetup.tsx

@@ -5,10 +5,7 @@ import { cn } from "@/lib/utils";
 import { Disabled } from "@/refresh-components/Disabled";
 import Text from "@/refresh-components/texts/Text";
 import SimpleTooltip from "@/refresh-components/SimpleTooltip";
-import {
-  LLMProviderName,
-  LLMProviderDescriptor,
-} from "@/app/admin/configuration/llm/interfaces";
+import { LLMProviderName, LLMProviderDescriptor } from "@/interfaces/llm";
 
 // Provider configurations
 export type ProviderKey = "anthropic" | "openai" | "openrouter";

web/src/app/craft/onboarding/constants.ts

@@ -19,13 +19,12 @@ const LLM_SELECTION_PRIORITY = [
 interface MinimalLlmProvider {
   name: string;
   provider: string;
-  default_model_name: string;
-  is_default_provider: boolean | null;
+  model_configurations: { name: string; is_visible: boolean }[];
 }
 
 /**
  * Get the best default LLM selection based on available providers.
- * Priority: Anthropic > OpenAI > OpenRouter > system default > first available
+ * Priority: Anthropic > OpenAI > OpenRouter > first available
  */
 export function getDefaultLlmSelection(
   llmProviders: MinimalLlmProvider[] | undefined
@@ -44,23 +43,16 @@ export function getDefaultLlmSelection(
     }
   }
 
-  // Fallback: use the default provider's default model
-  const defaultProvider = llmProviders.find((p) => p.is_default_provider);
-  if (defaultProvider) {
-    return {
-      providerName: defaultProvider.name,
-      provider: defaultProvider.provider,
-      modelName: defaultProvider.default_model_name,
-    };
-  }
-
-  // Final fallback: first available provider
+  // Fallback: first available provider, use its first visible model
   const firstProvider = llmProviders[0];
   if (firstProvider) {
+    const firstModel = firstProvider.model_configurations.find(
+      (m) => m.is_visible
+    );
     return {
       providerName: firstProvider.name,
       provider: firstProvider.provider,
-      modelName: firstProvider.default_model_name,
+      modelName: firstModel?.name ?? "",
     };
   }
 

web/src/app/craft/onboarding/hooks/useOnboardingModal.ts

@@ -2,8 +2,8 @@
 
 import { useCallback, useState, useMemo, useEffect } from "react";
 import { useUser } from "@/providers/UserProvider";
-import { useLLMProviders } from "@/lib/hooks/useLLMProviders";
-import { LLMProviderName } from "@/app/admin/configuration/llm/interfaces";
+import { useLLMProviders } from "@/hooks/useLLMProviders";
+import { LLMProviderName } from "@/interfaces/llm";
 import {
   OnboardingModalMode,
   OnboardingModalController,
@@ -18,9 +18,7 @@ import { useBuildSessionStore } from "@/app/craft/hooks/useBuildSessionStore";
 
 // Check if all 3 build mode providers are configured (anthropic, openai, openrouter)
 function checkAllProvidersConfigured(
-  llmProviders:
-    | import("@/app/admin/configuration/llm/interfaces").LLMProviderDescriptor[]
-    | undefined
+  llmProviders: import("@/interfaces/llm").LLMProviderDescriptor[] | undefined
 ): boolean {
   if (!llmProviders || llmProviders.length === 0) {
     return false;
@@ -35,9 +33,7 @@ function checkAllProvidersConfigured(
 
 // Check if at least one provider is configured
 function checkHasAnyProvider(
-  llmProviders:
-    | import("@/app/admin/configuration/llm/interfaces").LLMProviderDescriptor[]
-    | undefined
+  llmProviders: import("@/interfaces/llm").LLMProviderDescriptor[] | undefined
 ): boolean {
   return !!(llmProviders && llmProviders.length > 0);
 }

web/src/app/craft/onboarding/types.ts

@@ -1,4 +1,8 @@
 import { WorkArea, Level } from "./constants";
+import type {
+  LLMProviderDescriptor,
+  LLMProviderResponse,
+} from "@/interfaces/llm";
 
 export interface BuildUserInfo {
   firstName: string;
@@ -33,9 +37,7 @@ export interface OnboardingModalController {
   close: () => void;
 
   // Data needed for modal
-  llmProviders:
-    | import("@/app/admin/configuration/llm/interfaces").LLMProviderDescriptor[]
-    | undefined;
+  llmProviders: LLMProviderDescriptor[] | undefined;
   initialValues: {
     firstName: string;
     lastName: string;
@@ -54,7 +56,6 @@ export interface OnboardingModalController {
   completeUserInfo: (info: BuildUserInfo) => Promise<void>;
   completeLlmSetup: () => Promise<void>;
   refetchLlmProviders: () => Promise<
-    | import("@/app/admin/configuration/llm/interfaces").LLMProviderDescriptor[]
-    | undefined
+    LLMProviderResponse<LLMProviderDescriptor> | undefined
   >;
 }

web/src/app/craft/v1/configure/components/ComingSoonConnectors.tsx

@@ -3,7 +3,7 @@
 import { useState } from "react";
 import Card from "@/refresh-components/cards/Card";
 import Text from "@/refresh-components/texts/Text";
-import { LineItemLayout } from "@/layouts/general-layouts";
+import { Content } from "@opal/layouts";
 import Separator from "@/refresh-components/Separator";
 import { ValidSources } from "@/lib/types";
 import { getSourceMetadata } from "@/lib/sources";
@@ -99,14 +99,15 @@ export default function ComingSoonConnectors() {
           const card = (
             <div key={type} className="opacity-60">
               <Card variant="secondary">
-                <LineItemLayout
+                <Content
                   icon={
                     type === ValidSources.Imap
                       ? OutlookIcon
                       : sourceMetadata.icon
                   }
                   title={displayName}
-                  center
+                  sizePreset="main-ui"
+                  variant="body"
                 />
               </Card>
             </div>
@@ -118,7 +119,12 @@ export default function ComingSoonConnectors() {
               card,
               <div key="onedrive" className="opacity-60">
                 <Card variant="secondary">
-                  <LineItemLayout icon={OneDriveIcon} title="OneDrive" center />
+                  <Content
+                    icon={OneDriveIcon}
+                    title="OneDrive"
+                    sizePreset="main-ui"
+                    variant="body"
+                  />
                 </Card>
               </div>,
             ];
@@ -130,7 +136,12 @@ export default function ComingSoonConnectors() {
               card,
               <div key="box" className="opacity-60">
                 <Card variant="secondary">
-                  <LineItemLayout icon={BoxIcon} title="Box" center />
+                  <Content
+                    icon={BoxIcon}
+                    title="Box"
+                    sizePreset="main-ui"
+                    variant="body"
+                  />
                 </Card>
               </div>,
             ];
@@ -141,13 +152,23 @@ export default function ComingSoonConnectors() {
         {/* Enterprise/ERP */}
         <div className="opacity-60">
           <Card variant="secondary">
-            <LineItemLayout icon={ServiceNowIcon} title="ServiceNow" center />
+            <Content
+              icon={ServiceNowIcon}
+              title="ServiceNow"
+              sizePreset="main-ui"
+              variant="body"
+            />
           </Card>
         </div>
         {/* Project Management */}
         <div className="opacity-60">
           <Card variant="secondary">
-            <LineItemLayout icon={TrelloIcon} title="Trello" center />
+            <Content
+              icon={TrelloIcon}
+              title="Trello"
+              sizePreset="main-ui"
+              variant="body"
+            />
           </Card>
         </div>
       </div>

web/src/app/craft/v1/configure/components/ConnectorCard.tsx

@@ -4,8 +4,7 @@ import { useState } from "react";
 import Card from "@/refresh-components/cards/Card";
 import Popover from "@/refresh-components/Popover";
 import LineItem from "@/refresh-components/buttons/LineItem";
-import Text from "@/refresh-components/texts/Text";
-import { Section, LineItemLayout } from "@/layouts/general-layouts";
+import { ContentAction } from "@opal/layouts";
 import { ValidSources } from "@/lib/types";
 import { getSourceMetadata } from "@/lib/sources";
 import { SvgMoreHorizontal, SvgPlug, SvgSettings, SvgTrash } from "@opal/icons";
@@ -41,15 +40,6 @@ interface ConnectorCardProps {
   onDelete: () => void;
 }
 
-const STATUS_COLORS: Record<ConnectorStatus, string> = {
-  connected: "bg-status-success-05",
-  connected_with_errors: "bg-status-warning-05",
-  indexing: "bg-status-warning-05 animate-pulse",
-  error: "bg-status-error-05",
-  deleting: "bg-status-error-05 animate-pulse",
-  not_connected: "bg-background-neutral-03",
-};
-
 function getStatusText(status: ConnectorStatus, docsIndexed: number): string {
   switch (status) {
     case "connected":
@@ -72,29 +62,6 @@ function getStatusText(status: ConnectorStatus, docsIndexed: number): string {
   }
 }
 
-function StatusDescription({
-  status,
-  docsIndexed,
-}: {
-  status: ConnectorStatus;
-  docsIndexed: number;
-}) {
-  return (
-    <Section
-      flexDirection="row"
-      alignItems="center"
-      gap={0.375}
-      width="fit"
-      height="fit"
-    >
-      <div className={cn(STATUS_COLORS[status], "w-2 h-2 rounded-full")} />
-      <Text secondaryBody text03>
-        {getStatusText(status, docsIndexed)}
-      </Text>
-    </Section>
-  );
-}
-
 export default function ConnectorCard({
   connectorType,
   config,
@@ -180,17 +147,8 @@ export default function ConnectorCard({
   const cardVariant =
     isAlwaysConnected || isConnected ? "primary" : "secondary";
 
-  // Use custom description if provided, otherwise show status
-  const descriptionContent = customDescription ? (
-    <Text secondaryBody text03>
-      {customDescription}
-    </Text>
-  ) : (
-    <StatusDescription
-      status={status}
-      docsIndexed={config?.docs_indexed || 0}
-    />
-  );
+  const descriptionText =
+    customDescription ?? getStatusText(status, config?.docs_indexed || 0);
 
   return (
     <div
@@ -198,12 +156,13 @@ export default function ConnectorCard({
       onClick={handleCardClick}
     >
       <Card variant={cardVariant}>
-        <LineItemLayout
+        <ContentAction
           icon={sourceMetadata.icon}
           title={sourceMetadata.displayName}
-          description={descriptionContent}
+          description={descriptionText}
+          sizePreset="main-content"
+          variant="section"
           rightChildren={rightContent}
-          center
         />
       </Card>
     </div>

web/src/app/craft/v1/configure/page.tsx

@@ -46,7 +46,7 @@ import Switch from "@/refresh-components/inputs/Switch";
 import SimpleTooltip from "@/refresh-components/SimpleTooltip";
 import NotAllowedModal from "@/app/craft/onboarding/components/NotAllowedModal";
 import { useOnboarding } from "@/app/craft/onboarding/BuildOnboardingProvider";
-import { useLLMProviders } from "@/lib/hooks/useLLMProviders";
+import { useLLMProviders } from "@/hooks/useLLMProviders";
 import { useUser } from "@/providers/UserProvider";
 import { getProviderIcon } from "@/app/admin/configuration/llm/utils";
 import {

web/src/app/layout.tsx

@@ -150,7 +150,8 @@ export default async function RootLayout({
   // middleware returns 402 for all non-allowlisted API calls, preventing data
   // leakage. The user sees a brief loading state before being redirected.
   const content =
-    productGating === ApplicationStatus.GATED_ACCESS ? (
+    productGating === ApplicationStatus.GATED_ACCESS ||
+    productGating === ApplicationStatus.SEAT_LIMIT_EXCEEDED ? (
       <GatedContentWrapper>{children}</GatedContentWrapper>
     ) : (
       children

web/src/components/NonSelectableConnectors.tsx

@@ -1,6 +1,6 @@
 import { ConnectorStatus } from "@/lib/types";
 import { ConnectorTitle } from "@/components/admin/connectors/ConnectorTitle";
-import * as InputLayouts from "@/layouts/input-layouts";
+import { Content } from "@opal/layouts";
 import Text from "@/refresh-components/texts/Text";
 import { SvgLock } from "@opal/icons";
 interface NonSelectableConnectorsProps {
@@ -20,7 +20,12 @@ export const NonSelectableConnectors = ({
 
   return (
     <div className="mt-6 mb-4">
-      <InputLayouts.Title title={title} description={description} />
+      <Content
+        title={title}
+        description={description}
+        sizePreset="main-content"
+        variant="section"
+      />
 
       <div className="p-3 border border-dashed border-border-02 rounded-12 bg-background-neutral-01">
         <div className="mb-2 flex items-center gap-1.5">

web/src/components/admin/ClientLayout.tsx

@@ -27,6 +27,7 @@ const SETTINGS_LAYOUT_PREFIXES = [
   "/admin/document-index-migration",
   "/admin/discord-bot",
   "/admin/theme",
+  "/admin/configuration/llm",
 ];
 
 export function ClientLayout({

web/src/components/chat/FederatedOAuthModal.tsx

@@ -9,7 +9,7 @@ import { getSourceMetadata } from "@/lib/sources";
 import useFederatedOAuthStatus from "@/hooks/useFederatedOAuthStatus";
 import { SvgLink } from "@opal/icons";
 import { Card } from "@/refresh-components/cards";
-import { LineItemLayout } from "@/layouts/general-layouts";
+import { ContentAction } from "@opal/layouts";
 
 export interface FederatedConnectorOAuthStatus {
   federated_connector_id: number;
@@ -136,10 +136,12 @@ export default function FederatedOAuthModal() {
 
             return (
               <Card key={connector.federated_connector_id}>
-                <LineItemLayout
+                <ContentAction
                   icon={sourceMetadata.icon}
                   title={sourceMetadata.displayName}
                   description={sourceMetadata.category}
+                  sizePreset="main-content"
+                  variant="section"
                   rightChildren={
                     <Button
                       secondary
@@ -149,7 +151,6 @@ export default function FederatedOAuthModal() {
                       Connect
                     </Button>
                   }
-                  center
                 />
               </Card>
             );

web/src/components/chat/ProviderContext.tsx

@@ -2,7 +2,7 @@
 import {
   WellKnownLLMProviderDescriptor,
   LLMProviderDescriptor,
-} from "@/app/admin/configuration/llm/interfaces";
+} from "@/interfaces/llm";
 import React, {
   createContext,
   useContext,
@@ -11,9 +11,9 @@ import React, {
   useCallback,
 } from "react";
 import { useUser } from "@/providers/UserProvider";
-import { useLLMProviders } from "@/lib/hooks/useLLMProviders";
+import { useLLMProviders } from "@/hooks/useLLMProviders";
 import { useLLMProviderOptions } from "@/lib/hooks/useLLMProviderOptions";
-import { testDefaultProvider as testDefaultProviderSvc } from "@/lib/llm/svc";
+import { testDefaultProvider as testDefaultProviderSvc } from "@/lib/llmConfig/svc";
 
 interface ProviderContextType {
   shouldShowConfigurationNeeded: boolean;

web/src/components/errorPages/AccessRestrictedPage.tsx

@@ -9,10 +9,12 @@ import { logout } from "@/lib/user";
 import { loadStripe } from "@stripe/stripe-js";
 import { NEXT_PUBLIC_CLOUD_ENABLED } from "@/lib/constants";
 import { useLicense } from "@/hooks/useLicense";
+import { useSettingsContext } from "@/providers/SettingsProvider";
+import { ApplicationStatus } from "@/interfaces/settings";
 import Text from "@/refresh-components/texts/Text";
 import { SvgLock } from "@opal/icons";
 
-const linkClassName = "text-action-link-05 hover:text-action-link-06";
+const linkClassName = "text-action-link-05 hover:text-action-link-06 underline";
 
 const fetchStripePublishableKey = async (): Promise<string> => {
   const response = await fetch("/api/tenants/stripe-publishable-key");
@@ -40,15 +42,30 @@ export default function AccessRestricted() {
   const [isLoading, setIsLoading] = useState(false);
   const [error, setError] = useState<string | null>(null);
   const { data: license } = useLicense();
+  const settings = useSettingsContext();
 
+  const isSeatLimitExceeded =
+    settings.settings.application_status ===
+    ApplicationStatus.SEAT_LIMIT_EXCEEDED;
   const hadPreviousLicense = license?.has_license === true;
   const showRenewalMessage = NEXT_PUBLIC_CLOUD_ENABLED || hadPreviousLicense;
 
-  const initialModalMessage = showRenewalMessage
-    ? NEXT_PUBLIC_CLOUD_ENABLED
-      ? "Your access to Onyx has been temporarily suspended due to a lapse in your subscription."
-      : "Your access to Onyx has been temporarily suspended due to a lapse in your license."
-    : "An Enterprise license is required to use Onyx. Your data is protected and will be available once a license is activated.";
+  function getSeatLimitMessage() {
+    const { used_seats, seat_count } = settings.settings;
+    const counts =
+      used_seats != null && seat_count != null
+        ? ` (${used_seats} users / ${seat_count} seats)`
+        : "";
+    return `Your organization has exceeded its licensed seat count${counts}. Access is restricted until the number of users is reduced or your license is upgraded.`;
+  }
+
+  const initialModalMessage = isSeatLimitExceeded
+    ? getSeatLimitMessage()
+    : showRenewalMessage
+      ? NEXT_PUBLIC_CLOUD_ENABLED
+        ? "Your access to Onyx has been temporarily suspended due to a lapse in your subscription."
+        : "Your access to Onyx has been temporarily suspended due to a lapse in your license."
+      : "An Enterprise license is required to use Onyx. Your data is protected and will be available once a license is activated.";
 
   const handleResubscribe = async () => {
     setIsLoading(true);
@@ -80,7 +97,32 @@ export default function AccessRestricted() {
 
       <Text text03>{initialModalMessage}</Text>
 
-      {NEXT_PUBLIC_CLOUD_ENABLED ? (
+      {isSeatLimitExceeded ? (
+        <>
+          <Text text03>
+            If you are an administrator, you can manage users on the{" "}
+            <Link className={linkClassName} href="/admin/users">
+              User Management
+            </Link>{" "}
+            page or upgrade your license on the{" "}
+            <Link className={linkClassName} href="/admin/billing">
+              Admin Billing
+            </Link>{" "}
+            page.
+          </Text>
+
+          <div className="flex flex-row gap-2">
+            <Button
+              onClick={async () => {
+                await logout();
+                window.location.reload();
+              }}
+            >
+              Log out
+            </Button>
+          </div>
+        </>
+      ) : NEXT_PUBLIC_CLOUD_ENABLED ? (
         <>
           <Text text03>
             To reinstate your access and continue benefiting from Onyx&apos;s
@@ -127,7 +169,7 @@ export default function AccessRestricted() {
             sign up through Stripe or reach out to{" "}
             <a className={linkClassName} href="mailto:support@onyx.app">
               support@onyx.app
-            </a>
+            </a>{" "}
             for billing assistance.
           </Text>
 

web/src/components/llm/LLMSelector.tsx

@@ -1,8 +1,8 @@
 "use client";
 
 import { useMemo } from "react";
-import { parseLlmDescriptor, structureValue } from "@/lib/llm/utils";
-import { LLMProviderDescriptor } from "@/app/admin/configuration/llm/interfaces";
+import { parseLlmDescriptor, structureValue } from "@/lib/llmConfig/utils";
+import { DefaultModel, LLMProviderDescriptor } from "@/interfaces/llm";
 import { getProviderIcon } from "@/app/admin/configuration/llm/utils";
 import InputSelect from "@/refresh-components/inputs/InputSelect";
 import { createIcon } from "@/components/icons/icons";
@@ -23,6 +23,7 @@ export interface LLMSelectorProps {
   name?: string;
   userSettings?: boolean;
   llmProviders: LLMProviderDescriptor[];
+  defaultText?: DefaultModel | null;
   currentLlm: string | null;
   onSelect: (value: string | null) => void;
   requiresImageGeneration?: boolean;
@@ -33,6 +34,7 @@ export default function LLMSelector({
   name,
   userSettings,
   llmProviders,
+  defaultText,
   currentLlm,
   onSelect,
   requiresImageGeneration,
@@ -139,11 +141,11 @@ export default function LLMSelector({
     });
   }, [llmOptions]);
 
-  const defaultProvider = llmProviders.find(
-    (llmProvider) => llmProvider.is_default_provider
-  );
+  const defaultProvider = defaultText
+    ? llmProviders.find((p) => p.id === defaultText.provider_id)
+    : undefined;
 
-  const defaultModelName = defaultProvider?.default_model_name;
+  const defaultModelName = defaultText?.model_name;
   const defaultModelConfig = defaultProvider?.model_configurations.find(
     (m) => m.name === defaultModelName
   );

web/src/components/modals/CreateProjectModal.tsx

@@ -3,6 +3,7 @@
 import { useState, useEffect } from "react";
 import Button from "@/refresh-components/buttons/Button";
 import { useProjectsContext } from "@/providers/ProjectsContext";
+import { useKeyPress } from "@/hooks/useKeyPress";
 import * as InputLayouts from "@/layouts/input-layouts";
 import { useAppRouter } from "@/hooks/appNavigation";
 import { useModal } from "@/refresh-components/contexts/ModalContext";
@@ -41,6 +42,8 @@ export default function CreateProjectModal({
     }
   }
 
+  useKeyPress(handleSubmit, "Enter");
+
   return (
     <>
       <Modal open={modal.isOpen} onOpenChange={modal.toggle}>

web/src/hooks/useChatController.ts

@@ -42,7 +42,7 @@ import {
   getFinalLLM,
   modelSupportsImageInput,
   structureValue,
-} from "@/lib/llm/utils";
+} from "@/lib/llmConfig/utils";
 import {
   CurrentMessageFIFO,
   updateCurrentMessageFIFO,

web/src/hooks/useLLMProviders.ts

@@ -0,0 +1,150 @@
+"use client";
+
+import useSWR from "swr";
+import { errorHandlingFetcher } from "@/lib/fetcher";
+import {
+  LLMProviderDescriptor,
+  LLMProviderResponse,
+  LLMProviderView,
+  WellKnownLLMProviderDescriptor,
+} from "@/interfaces/llm";
+import { LLM_PROVIDERS_ADMIN_URL } from "@/lib/llmConfig/constants";
+
+/**
+ * Fetches configured LLM providers accessible to the current user.
+ *
+ * Hits the **non-admin** endpoints which return `LLMProviderDescriptor`
+ * (no `id` or sensitive fields like `api_key`). Use this hook in
+ * user-facing UI (chat, popovers, onboarding) where you need the list
+ * of providers and their visible models but don't need admin-level details.
+ *
+ * The backend wraps the provider list in an `LLMProviderResponse` envelope
+ * that also carries the global default text and vision models. This hook
+ * unwraps `.providers` for convenience while still exposing the defaults.
+ *
+ * **Endpoints:**
+ * - No `personaId` → `GET /api/llm/provider`
+ *   Returns all public providers plus restricted providers the user can
+ *   access via group membership.
+ * - With `personaId` → `GET /api/llm/persona/{personaId}/providers`
+ *   Returns providers scoped to a specific persona, respecting RBAC
+ *   restrictions. Use this when displaying model options for a particular
+ *   assistant.
+ *
+ * @param personaId - Optional persona ID for RBAC-scoped providers.
+ *
+ * @returns
+ * - `llmProviders` — The array of provider descriptors, or `undefined`
+ *    while loading.
+ * - `defaultText` — The global (or persona-overridden) default text model.
+ * - `defaultVision` — The global (or persona-overridden) default vision model.
+ * - `isLoading` — `true` until the first successful response or error.
+ * - `error` — The SWR error object, if any.
+ * - `refetch` — SWR `mutate` function to trigger a revalidation.
+ */
+export function useLLMProviders(personaId?: number) {
+  const url =
+    personaId !== undefined
+      ? `/api/llm/persona/${personaId}/providers`
+      : "/api/llm/provider";
+
+  const { data, error, mutate } = useSWR<
+    LLMProviderResponse<LLMProviderDescriptor>
+  >(url, errorHandlingFetcher, {
+    revalidateOnFocus: false,
+    dedupingInterval: 60000,
+  });
+
+  return {
+    llmProviders: data?.providers,
+    defaultText: data?.default_text ?? null,
+    defaultVision: data?.default_vision ?? null,
+    isLoading: !error && !data,
+    error,
+    refetch: mutate,
+  };
+}
+
+/**
+ * Fetches configured LLM providers via the **admin** endpoint.
+ *
+ * Hits `GET /api/admin/llm/provider` which returns `LLMProviderView` —
+ * the full provider object including `id`, `api_key` (masked),
+ * group/persona assignments, and all other admin-visible fields.
+ *
+ * Use this hook on admin pages (e.g. the LLM Configuration page) where
+ * you need provider IDs for mutations (setting defaults, editing, deleting)
+ * or need to display admin-only metadata. **Do not use in user-facing UI**
+ * — use `useLLMProviders` instead.
+ *
+ * @returns
+ * - `llmProviders` — The array of full provider views, or `undefined`
+ *    while loading.
+ * - `defaultText` — The global default text model.
+ * - `defaultVision` — The global default vision model.
+ * - `isLoading` — `true` until the first successful response or error.
+ * - `error` — The SWR error object, if any.
+ * - `refetch` — SWR `mutate` function to trigger a revalidation.
+ */
+export function useAdminLLMProviders() {
+  const { data, error, mutate } = useSWR<LLMProviderResponse<LLMProviderView>>(
+    LLM_PROVIDERS_ADMIN_URL,
+    errorHandlingFetcher,
+    {
+      revalidateOnFocus: false,
+      dedupingInterval: 60000,
+    }
+  );
+
+  return {
+    llmProviders: data?.providers,
+    defaultText: data?.default_text ?? null,
+    defaultVision: data?.default_vision ?? null,
+    isLoading: !error && !data,
+    error,
+    refetch: mutate,
+  };
+}
+
+/**
+ * Fetches the catalog of well-known (built-in) LLM providers.
+ *
+ * Hits `GET /api/admin/llm/built-in/options` which returns the static
+ * list of provider descriptors that Onyx ships with out of the box
+ * (OpenAI, Anthropic, Vertex AI, Bedrock, Azure, Ollama, OpenRouter,
+ * etc.). Each descriptor includes the provider's known models and the
+ * recommended default model.
+ *
+ * Used primarily on the LLM Configuration page and onboarding flows
+ * to show which providers are available to set up, and to pre-populate
+ * model lists before the user has entered credentials.
+ *
+ * @returns
+ * - `wellKnownLLMProviders` — The array of built-in provider descriptors,
+ *    or `null` while loading.
+ * - `isLoading` — `true` until the first successful response or error.
+ * - `error` — The SWR error object, if any.
+ * - `mutate` — SWR `mutate` function to trigger a revalidation.
+ */
+export function useWellKnownLLMProviders() {
+  const {
+    data: wellKnownLLMProviders,
+    error,
+    isLoading,
+    mutate,
+  } = useSWR<WellKnownLLMProviderDescriptor[]>(
+    "/api/admin/llm/built-in/options",
+    errorHandlingFetcher,
+    {
+      revalidateOnFocus: false,
+      dedupingInterval: 60000,
+    }
+  );
+
+  return {
+    wellKnownLLMProviders: wellKnownLLMProviders ?? null,
+    isLoading,
+    error,
+    mutate,
+  };
+}

web/src/interfaces/llm.ts

@@ -13,8 +13,8 @@ export interface ModelConfiguration {
   name: string;
   is_visible: boolean;
   max_input_tokens: number | null;
-  supports_image_input: boolean | null;
-  supports_reasoning?: boolean;
+  supports_image_input: boolean;
+  supports_reasoning: boolean;
   display_name?: string;
   provider_display_name?: string;
   vendor?: string;
@@ -30,7 +30,6 @@ export interface SimpleKnownModel {
 export interface WellKnownLLMProviderDescriptor {
   name: string;
   known_models: ModelConfiguration[];
-
   recommended_default_model: SimpleKnownModel | null;
 }
 
@@ -40,44 +39,31 @@ export interface LLMModelDescriptor {
   maxTokens: number;
 }
 
-export interface LLMProvider {
+export interface LLMProviderView {
+  id: number;
   name: string;
   provider: string;
   api_key: string | null;
   api_base: string | null;
   api_version: string | null;
   custom_config: { [key: string]: string } | null;
-  default_model_name: string;
   is_public: boolean;
   is_auto_mode: boolean;
   groups: number[];
   personas: number[];
   deployment_name: string | null;
-  default_vision_model: string | null;
-  is_default_vision_provider: boolean | null;
   model_configurations: ModelConfiguration[];
 }
 
-export interface LLMProviderView extends LLMProvider {
-  id: number;
-  is_default_provider: boolean | null;
-}
-
 export interface VisionProvider extends LLMProviderView {
   vision_models: string[];
 }
 
 export interface LLMProviderDescriptor {
+  id: number;
   name: string;
   provider: string;
-  provider_display_name?: string;
-  default_model_name: string;
-  is_default_provider: boolean | null;
-  is_default_vision_provider?: boolean | null;
-  default_vision_model?: string | null;
-  is_public?: boolean;
-  groups?: number[];
-  personas?: number[];
+  provider_display_name: string;
   model_configurations: ModelConfiguration[];
 }
 
@@ -102,9 +88,22 @@ export interface BedrockModelResponse {
   supports_image_input: boolean;
 }
 
+export interface DefaultModel {
+  provider_id: number;
+  model_name: string;
+}
+
+export interface LLMProviderResponse<T> {
+  providers: T[];
+  default_text: DefaultModel | null;
+  default_vision: DefaultModel | null;
+}
+
 export interface LLMProviderFormProps {
   existingLlmProvider?: LLMProviderView;
   shouldMarkAsDefault?: boolean;
+  open?: boolean;
+  onOpenChange?: (open: boolean) => void;
 }
 
 // Param types for model fetching functions - use snake_case to match API structure

web/src/interfaces/settings.ts

@@ -2,6 +2,7 @@ export enum ApplicationStatus {
   PAYMENT_REMINDER = "payment_reminder",
   GATED_ACCESS = "gated_access",
   ACTIVE = "active",
+  SEAT_LIMIT_EXCEEDED = "seat_limit_exceeded",
 }
 
 export enum QueryHistoryType {
@@ -49,6 +50,10 @@ export interface Settings {
   // True when user has a valid license, False for community edition
   ee_features_enabled?: boolean;
 
+  // Seat usage - populated when seat limit is exceeded
+  seat_count?: number | null;
+  used_seats?: number | null;
+
   // OpenSearch migration
   opensearch_indexing_enabled?: boolean;
 

web/src/layouts/actions-layouts.tsx

@@ -56,7 +56,7 @@
 import React, { HtmlHTMLAttributes } from "react";
 import type { IconProps } from "@opal/types";
 import { WithoutStyles } from "@/types";
-import { LineItemLayout } from "@/layouts/general-layouts";
+import { ContentAction } from "@opal/layouts";
 import * as ExpandableCard from "@/layouts/expandable-card-layouts";
 import { Card } from "@/refresh-components/cards";
 import Label from "@/refresh-components/form/Label";
@@ -119,11 +119,14 @@ function ActionsHeader({
       <div className="flex flex-col gap-2 pt-4 pb-2">
         <div className="px-4">
           <Label name={name}>
-            <LineItemLayout
+            <ContentAction
               icon={Icon}
               title={title}
               description={description}
+              sizePreset="section"
+              variant="section"
               rightChildren={rightChildren}
+              paddingVariant="fit"
             />
           </Label>
         </div>
@@ -237,12 +240,14 @@ function ActionsTool({
   return (
     <Card padding={0.75} variant={disabled ? "disabled" : undefined}>
       <Label name={name} disabled={disabled}>
-        <LineItemLayout
+        <ContentAction
           icon={icon}
           title={title}
           description={description}
+          sizePreset="main-ui"
+          variant="section"
           rightChildren={rightChildren}
-          variant="secondary"
+          paddingVariant="fit"
         />
       </Label>
     </Card>

web/src/layouts/app-layouts.tsx

@@ -48,7 +48,6 @@ import { PopoverSearchInput } from "@/sections/sidebar/ChatButton";
 import SimplePopover from "@/refresh-components/SimplePopover";
 import { Interactive } from "@opal/core";
 import { Button, OpenButton } from "@opal/components";
-import { LineItemLayout } from "@/layouts/general-layouts";
 import { useAppSidebarContext } from "@/providers/AppSidebarProvider";
 import useScreenSize from "@/hooks/useScreenSize";
 import {

web/src/layouts/general-layouts.tsx

@@ -2,6 +2,7 @@ import { cn } from "@/lib/utils";
 import Text from "@/refresh-components/texts/Text";
 import Truncated from "@/refresh-components/texts/Truncated";
 import { WithoutStyles } from "@/types";
+import { Content } from "@opal/layouts";
 import { IconProps } from "@opal/types";
 import React from "react";
 
@@ -160,152 +161,6 @@ function Section({
   );
 }
 
-/**
- * LineItemLayout - A layout for icon + title + description rows
- *
- * Structure:
- *   Flexbox Row [
- *     Grid [
- *       [Icon] [Title      ]
- *       [    ] [Description]
- *     ],
- *     rightChildren
- *   ]
- *
- * - Icon column auto-sizes to icon width
- * - Icon vertically centers with title
- * - Description aligns with title's left edge (both in grid column 2)
- * - rightChildren is outside the grid, in the outer flexbox
- *
- * Variants:
- * - `primary`: Standard size (20px icon) with emphasized text. The default for prominent list items.
- * - `secondary`: Compact size (16px icon) with standard text. Use for denser lists or nested items.
- * - `tertiary`: Compact size (16px icon) with standard text. Use for less prominent items in tight layouts.
- * - `tertiary-muted`: Compact size (16px icon) with muted text styling. Use for de-emphasized or secondary information.
- * - `mini`: Smallest size (12px icon) with muted secondary text. Use for metadata labels (e.g., owner, action count).
- *
- * @param icon - Optional icon component to display on the left
- * @param title - The main title text (required)
- * @param description - Optional description content below the title (string or ReactNode)
- * @param rightChildren - Optional content to render on the right side
- * @param variant - Visual variant. Default: "primary"
- * @param strikethrough - If true, applies line-through style to title. Default: false
- * @param loading - If true, renders skeleton placeholders instead of content. Default: false
- * @param center - If true, vertically centers items; otherwise aligns to start. Default: false
- */
-type LineItemLayoutVariant =
-  | "primary"
-  | "secondary"
-  | "tertiary"
-  | "tertiary-muted"
-  | "mini";
-export interface LineItemLayoutProps {
-  icon?: React.FunctionComponent<IconProps>;
-  title: string;
-  description?: React.ReactNode;
-  middleText?: string;
-  rightChildren?: React.ReactNode;
-
-  variant?: LineItemLayoutVariant;
-  width?: Length;
-  strikethrough?: boolean;
-  loading?: boolean;
-  center?: boolean;
-  reducedPadding?: boolean;
-}
-function LineItemLayout({
-  icon: Icon,
-  title,
-  description,
-  middleText,
-  rightChildren,
-
-  variant = "primary",
-  width,
-  strikethrough,
-  loading,
-  center,
-  reducedPadding,
-}: LineItemLayoutProps) {
-  // Derive styling from variant
-  const isMini = variant === "mini";
-  const isCompact =
-    variant === "secondary" ||
-    variant === "tertiary" ||
-    variant === "tertiary-muted";
-  const isMuted = variant === "tertiary-muted" || isMini;
-
-  // Determine icon size: mini=12px, compact=16px, primary=20px
-  const iconSize = isMini ? 12 : isCompact ? 16 : 20;
-
-  // Determine gap: mini=0.25rem, others=1.5rem
-  const gap = isMini ? 0.25 : 1.5;
-
-  return (
-    <Section
-      flexDirection="row"
-      justifyContent="between"
-      alignItems={center || isMini ? "center" : "start"}
-      gap={gap}
-      width={width}
-    >
-      <div
-        className="line-item-layout"
-        data-variant={variant}
-        data-has-icon={Icon ? "true" : undefined}
-        data-loading={loading ? "true" : undefined}
-        data-strikethrough={strikethrough ? "true" : undefined}
-        data-reduced-padding={reducedPadding ? "true" : undefined}
-      >
-        {/* Row 1: Icon, Title */}
-        {Icon && <Icon size={iconSize} className="line-item-layout-icon" />}
-        {loading ? (
-          <div className="line-item-layout-skeleton-title" />
-        ) : (
-          <Truncated
-            mainContentEmphasis={!isCompact && !isMini}
-            secondaryBody={isMini}
-            mainUiAction={variant === "secondary"}
-            text03={isMuted}
-            className="line-item-layout-title"
-          >
-            {title}
-          </Truncated>
-        )}
-
-        {/* Row 2: Description (column 2, or column 1 if no icon) */}
-        {loading && description ? (
-          <div className="line-item-layout-skeleton-description" />
-        ) : description ? (
-          <div className="line-item-layout-description">
-            {typeof description === "string" ? (
-              <Text secondaryBody text03>
-                {description}
-              </Text>
-            ) : (
-              description
-            )}
-          </div>
-        ) : undefined}
-      </div>
-
-      {!loading && middleText && (
-        <div className="flex-1">
-          <Truncated text03 secondaryBody>
-            {middleText}
-          </Truncated>
-        </div>
-      )}
-
-      {loading && rightChildren ? (
-        <div className="line-item-layout-skeleton-right" />
-      ) : rightChildren ? (
-        <div className="flex-shrink-0">{rightChildren}</div>
-      ) : undefined}
-    </Section>
-  );
-}
-
 export interface AttachmentItemLayoutProps {
   title: string;
   description: string;
@@ -329,18 +184,29 @@ function AttachmentItemLayout({
           </div>
         </Section>
       </div>
-      <LineItemLayout
-        title={title}
-        description={description}
-        middleText={middleText}
-        rightChildren={
-          rightChildren ? (
-            <div className="px-1">{rightChildren}</div>
-          ) : undefined
-        }
-        center
-        variant="secondary"
-      />
+      <Section
+        flexDirection="row"
+        justifyContent="between"
+        alignItems="center"
+        gap={1.5}
+      >
+        <Content
+          title={title}
+          description={description}
+          sizePreset="main-ui"
+          variant="section"
+        />
+        {middleText && (
+          <div className="flex-1">
+            <Truncated text03 secondaryBody>
+              {middleText}
+            </Truncated>
+          </div>
+        )}
+        {rightChildren && (
+          <div className="flex-shrink-0 px-1">{rightChildren}</div>
+        )}
+      </Section>
     </Section>
   );
 }
@@ -413,4 +279,4 @@ function CardItemLayout({
     </div>
   );
 }
-export { Section, LineItemLayout, CardItemLayout, AttachmentItemLayout };
+export { Section, CardItemLayout, AttachmentItemLayout };

web/src/layouts/input-layouts.tsx

@@ -4,13 +4,18 @@ import Text from "@/refresh-components/texts/Text";
 import { SvgXOctagon, SvgAlertCircle } from "@opal/icons";
 import { useField, useFormikContext } from "formik";
 import { Section } from "@/layouts/general-layouts";
+import { Content } from "@opal/layouts";
 import Label from "@/refresh-components/form/Label";
 
-interface OrientationLayoutProps extends TitleLayoutProps {
+interface OrientationLayoutProps {
   name?: string;
   disabled?: boolean;
   nonInteractive?: boolean;
   children?: React.ReactNode;
+  title: string;
+  description?: string;
+  optional?: boolean;
+  sizePreset?: "main-content" | "main-ui";
 }
 
 /**
@@ -44,11 +49,20 @@ function VerticalInputLayout({
   nonInteractive,
   children,
   subDescription,
-  ...titleLayoutProps
+  title,
+  description,
+  optional,
+  sizePreset = "main-content",
 }: VerticalLayoutProps) {
   const content = (
     <Section gap={0.25} alignItems="start">
-      <TitleLayout {...titleLayoutProps} />
+      <Content
+        title={title}
+        description={description}
+        optional={optional}
+        sizePreset={sizePreset}
+        variant="section"
+      />
       {children}
       {name && <ErrorLayout name={name} />}
       {subDescription && (
@@ -110,7 +124,10 @@ function HorizontalInputLayout({
   nonInteractive,
   children,
   center,
-  ...titleLayoutProps
+  title,
+  description,
+  optional,
+  sizePreset = "main-content",
 }: HorizontalLayoutProps) {
   const content = (
     <Section gap={0.25} alignItems="start">
@@ -120,7 +137,13 @@ function HorizontalInputLayout({
         alignItems={center ? "center" : "start"}
       >
         <div className="flex flex-col flex-1 self-stretch">
-          <TitleLayout {...titleLayoutProps} />
+          <Content
+            title={title}
+            description={description}
+            optional={optional}
+            sizePreset={sizePreset}
+            variant="section"
+          />
         </div>
         <div className="flex flex-col items-end">{children}</div>
       </Section>
@@ -136,80 +159,6 @@ function HorizontalInputLayout({
   );
 }
 
-/**
- * TitleLayout - A reusable title/description component for form fields
- *
- * Renders a title with an optional description and "Optional" indicator.
- * This is a pure presentational component — it does not render a `<label>`
- * element. Label semantics are handled by the parent orientation layout
- * (Vertical/Horizontal) or by the caller.
- *
- * Exported as `Title` for convenient usage.
- *
- * @param title - The main label text
- * @param description - Additional helper text shown below the title
- * @param optional - Whether to show "(Optional)" indicator
- * @param center - If true, centers the title and description text. Default: false
- *
- * @example
- * ```tsx
- * import { Title } from "@/layouts/input-layouts";
- *
- * <Title
- *   name="username"
- *   title="Username"
- *   description="Choose a unique username"
- *   optional
- * />
- * ```
- */
-type TitleLayoutVariants = "primary" | "secondary";
-export interface TitleLayoutProps {
-  title: string;
-  description?: string;
-  optional?: boolean;
-  center?: boolean;
-  variant?: TitleLayoutVariants;
-}
-function TitleLayout({
-  title,
-  description,
-  optional,
-  center,
-  variant = "primary",
-}: TitleLayoutProps) {
-  return (
-    <Section gap={0} height="fit">
-      <Section
-        flexDirection="row"
-        justifyContent={center ? "center" : "start"}
-        gap={0.25}
-      >
-        <Text
-          mainContentEmphasis={variant === "primary"}
-          mainUiAction={variant === "secondary"}
-          text04
-        >
-          {title}
-        </Text>
-        {optional && (
-          <Text text03 mainContentMuted>
-            (Optional)
-          </Text>
-        )}
-      </Section>
-
-      {description && (
-        <Section alignItems={center ? "center" : "start"}>
-          <Text secondaryBody text03>
-            {description}
-          </Text>
-        </Section>
-      )}
-    </Section>
-  );
-}
-
 /**
  * ErrorLayout - Displays Formik field validation errors
  *
@@ -280,7 +229,6 @@ function ErrorTextLayout({ children, type = "error" }: ErrorTextLayoutProps) {
 export {
   VerticalInputLayout as Vertical,
   HorizontalInputLayout as Horizontal,
-  TitleLayout as Title,
   ErrorLayout as Error,
   ErrorTextLayout,
 };

web/src/lib/billing/interfaces.ts

@@ -18,7 +18,8 @@ export type ApplicationStatus =
   | "active"
   | "payment_reminder"
   | "gated_access"
-  | "expired";
+  | "expired"
+  | "seat_limit_exceeded";
 
 /**
  * Billing status from Stripe subscription.

web/src/lib/hooks.llmResolver.test.ts

@@ -2,8 +2,8 @@ import {
   getDefaultLlmDescriptor,
   getValidLlmDescriptorForProviders,
 } from "@/lib/hooks";
-import { structureValue } from "@/lib/llm/utils";
-import { LLMProviderDescriptor } from "@/app/admin/configuration/llm/interfaces";
+import { structureValue } from "@/lib/llmConfig/utils";
+import { LLMProviderDescriptor } from "@/interfaces/llm";
 import { makeProvider } from "@tests/setup/llmProviderTestUtils";
 
 describe("LLM resolver helpers", () => {
@@ -11,29 +11,30 @@ describe("LLM resolver helpers", () => {
     const sharedModel = "shared-runtime-model";
     const providers: LLMProviderDescriptor[] = [
       makeProvider({
+        id: 1,
         name: "OpenAI Provider",
         provider: "openai",
-        default_model_name: sharedModel,
-        is_default_provider: true,
         model_configurations: [
           {
             name: sharedModel,
             is_visible: true,
             max_input_tokens: null,
             supports_image_input: false,
+            supports_reasoning: false,
           },
         ],
       }),
       makeProvider({
+        id: 2,
         name: "Anthropic Provider",
         provider: "anthropic",
-        default_model_name: sharedModel,
         model_configurations: [
           {
             name: sharedModel,
             is_visible: true,
             max_input_tokens: null,
             supports_image_input: false,
+            supports_reasoning: false,
           },
         ],
       }),
@@ -54,29 +55,30 @@ describe("LLM resolver helpers", () => {
   test("falls back to default provider when model is unavailable", () => {
     const providers: LLMProviderDescriptor[] = [
       makeProvider({
+        id: 10,
         name: "Default OpenAI",
         provider: "openai",
-        default_model_name: "gpt-4o-mini",
-        is_default_provider: true,
         model_configurations: [
           {
             name: "gpt-4o-mini",
             is_visible: true,
             max_input_tokens: null,
             supports_image_input: true,
+            supports_reasoning: false,
           },
         ],
       }),
       makeProvider({
+        id: 20,
         name: "Anthropic Backup",
         provider: "anthropic",
-        default_model_name: "claude-3-5-sonnet",
         model_configurations: [
           {
             name: "claude-3-5-sonnet",
             is_visible: true,
             max_input_tokens: null,
             supports_image_input: true,
+            supports_reasoning: false,
           },
         ],
       }),
@@ -97,30 +99,30 @@ describe("LLM resolver helpers", () => {
   test("uses first provider with models when no explicit default exists", () => {
     const providers: LLMProviderDescriptor[] = [
       makeProvider({
+        id: 30,
         name: "First Provider",
         provider: "openai",
-        default_model_name: "gpt-first",
-        is_default_provider: false,
         model_configurations: [
           {
             name: "gpt-first",
             is_visible: true,
             max_input_tokens: null,
             supports_image_input: false,
+            supports_reasoning: false,
           },
         ],
       }),
       makeProvider({
+        id: 40,
         name: "Second Provider",
         provider: "anthropic",
-        default_model_name: "claude-second",
-        is_default_provider: false,
         model_configurations: [
           {
             name: "claude-second",
             is_visible: true,
             max_input_tokens: null,
             supports_image_input: false,
+            supports_reasoning: false,
           },
         ],
       }),

web/src/lib/hooks.ts

@@ -23,23 +23,22 @@ import {
 } from "react";
 import { DateRangePickerValue } from "@/components/dateRangeSelectors/AdminDateRangeSelector";
 import { SourceMetadata } from "./search/interfaces";
-import { parseLlmDescriptor } from "./llm/utils";
+import { parseLlmDescriptor } from "./llmConfig/utils";
 import { ChatSession } from "@/app/app/interfaces";
-import { AllUsersResponse } from "./types";
 import { Credential } from "./connectors/credentials";
 import { SettingsContext } from "@/providers/SettingsProvider";
 import {
   MinimalPersonaSnapshot,
   PersonaLabel,
 } from "@/app/admin/assistants/interfaces";
-import { LLMProviderDescriptor } from "@/app/admin/configuration/llm/interfaces";
+import { DefaultModel, LLMProviderDescriptor } from "@/interfaces/llm";
 import { isAnthropic } from "@/app/admin/configuration/llm/utils";
 import { getSourceMetadataForSources } from "./sources";
 import { AuthType, NEXT_PUBLIC_CLOUD_ENABLED } from "./constants";
 import { useUser } from "@/providers/UserProvider";
 import { SEARCH_TOOL_ID } from "@/app/app/components/tools/constants";
 import { updateTemperatureOverrideForChatSession } from "@/app/app/services/lib";
-import { useLLMProviders } from "./hooks/useLLMProviders";
+import { useLLMProviders } from "@/hooks/useLLMProviders";
 
 const CREDENTIAL_URL = "/api/manage/admin/credential";
 
@@ -533,26 +532,31 @@ providing appropriate defaults for new conversations based on the available tool
 */
 
 export function getDefaultLlmDescriptor(
-  llmProviders: LLMProviderDescriptor[]
+  llmProviders: LLMProviderDescriptor[],
+  defaultText?: DefaultModel | null
 ): LlmDescriptor | null {
-  const defaultProvider = llmProviders.find(
-    (provider) => provider.is_default_provider
-  );
-  if (defaultProvider) {
-    return {
-      name: defaultProvider.name,
-      provider: defaultProvider.provider,
-      modelName: defaultProvider.default_model_name,
-    };
+  if (defaultText) {
+    const provider = llmProviders.find((p) => p.id === defaultText.provider_id);
+    if (provider) {
+      return {
+        name: provider.name,
+        provider: provider.provider,
+        modelName: defaultText.model_name,
+      };
+    }
   }
+  // Fallback: first provider with visible models
   const firstLlmProvider = llmProviders.find(
     (provider) => provider.model_configurations.length > 0
   );
   if (firstLlmProvider) {
+    const firstModel = firstLlmProvider.model_configurations.find(
+      (m) => m.is_visible
+    );
     return {
       name: firstLlmProvider.name,
       provider: firstLlmProvider.provider,
-      modelName: firstLlmProvider.default_model_name,
+      modelName: firstModel?.name ?? "",
     };
   }
   return null;
@@ -637,19 +641,25 @@ export function useLlmManager(
 
   // Get all user-accessible providers via SWR (general providers - no persona filter)
   // This includes public + all restricted providers user can access via groups
-  const { llmProviders: allUserProviders, isLoading: isLoadingAllProviders } =
-    useLLMProviders();
+  const {
+    llmProviders: allUserProviders,
+    defaultText: allUserDefaultText,
+    isLoading: isLoadingAllProviders,
+  } = useLLMProviders();
   // Fetch persona-specific providers to enforce RBAC restrictions per assistant
   // Only fetch if we have an assistant selected
   const personaId =
     liveAssistant?.id !== undefined ? liveAssistant.id : undefined;
   const {
     llmProviders: personaProviders,
+    defaultText: personaDefaultText,
     isLoading: isLoadingPersonaProviders,
   } = useLLMProviders(personaId);
 
   const llmProviders =
     personaProviders !== undefined ? personaProviders : allUserProviders;
+  const defaultText =
+    personaProviders !== undefined ? personaDefaultText : allUserDefaultText;
 
   const [userHasManuallyOverriddenLLM, setUserHasManuallyOverriddenLLM] =
     useState(false);
@@ -708,7 +718,7 @@ export function useLlmManager(
       } else if (user?.preferences?.default_model) {
         setCurrentLlm(getValidLlmDescriptor(user.preferences.default_model));
       } else {
-        const defaultLlm = getDefaultLlmDescriptor(llmProviders);
+        const defaultLlm = getDefaultLlmDescriptor(llmProviders, defaultText);
         if (defaultLlm) {
           setCurrentLlm(defaultLlm);
         }

web/src/lib/hooks/useLLMProviderOptions.ts

@@ -1,5 +1,5 @@
 import useSWR from "swr";
-import { WellKnownLLMProviderDescriptor } from "@/app/admin/configuration/llm/interfaces";
+import { WellKnownLLMProviderDescriptor } from "@/interfaces/llm";
 import { errorHandlingFetcher } from "@/lib/fetcher";
 
 export function useLLMProviderOptions() {

web/src/lib/hooks/useLLMProviders.test.ts

@@ -1,5 +1,5 @@
 import useSWR from "swr";
-import { useLLMProviders } from "./useLLMProviders";
+import { useLLMProviders } from "@/hooks/useLLMProviders";
 import { errorHandlingFetcher } from "@/lib/fetcher";
 
 jest.mock("swr", () => ({
@@ -45,7 +45,7 @@ describe("useLLMProviders", () => {
     const mockMutate = jest.fn();
     const providers = [{ name: "Persona Provider" }];
     mockUseSWR.mockReturnValue({
-      data: providers,
+      data: { providers, default_text: null, default_vision: null },
       error: undefined,
       mutate: mockMutate,
       isValidating: false,

web/src/lib/hooks/useLLMProviders.ts

@@ -1,30 +0,0 @@
-import useSWR from "swr";
-import { LLMProviderDescriptor } from "@/app/admin/configuration/llm/interfaces";
-import { errorHandlingFetcher } from "@/lib/fetcher";
-
-export function useLLMProviders(personaId?: number) {
-  // personaId can be:
-  // - undefined: public providers only (/api/llm/provider)
-  // - number (personaId): persona-specific providers with RBAC enforcement
-
-  const url =
-    typeof personaId === "number"
-      ? `/api/llm/persona/${personaId}/providers`
-      : "/api/llm/provider";
-
-  const { data, error, mutate } = useSWR<LLMProviderDescriptor[] | undefined>(
-    url,
-    errorHandlingFetcher,
-    {
-      revalidateOnFocus: false, // Cache aggressively for performance
-      dedupingInterval: 60000, // Dedupe requests within 1 minute
-    }
-  );
-
-  return {
-    llmProviders: data,
-    isLoading: !error && !data,
-    error,
-    refetch: mutate,
-  };
-}

web/src/lib/llm/fetchLLMs.ts

@@ -1,10 +0,0 @@
-import { LLMProviderDescriptor } from "@/app/admin/configuration/llm/interfaces";
-import { fetchSS } from "../utilsSS";
-
-export async function fetchLLMProvidersSS() {
-  const response = await fetchSS("/llm/provider");
-  if (response.ok) {
-    return (await response.json()) as LLMProviderDescriptor[];
-  }
-  return [];
-}

web/src/lib/llm/svc.ts

@@ -1,23 +0,0 @@
-/**
- * LLM action functions for mutations.
- *
- * These are async functions for one-off actions that don't need SWR caching.
- *
- * Endpoints:
- * - /api/admin/llm/test/default - Test the default LLM provider connection
- */
-
-/**
- * Test the default LLM provider.
- * Returns true if the default provider is configured and working, false otherwise.
- */
-export async function testDefaultProvider(): Promise<boolean> {
-  try {
-    const response = await fetch("/api/admin/llm/test/default", {
-      method: "POST",
-    });
-    return response?.ok || false;
-  } catch {
-    return false;
-  }
-}

web/src/lib/llmConfig/constants.ts

@@ -1,4 +1,5 @@
-export const LLM_PROVIDERS_ADMIN_URL = "/api/admin/llm/provider";
+export const LLM_ADMIN_URL = "/api/admin/llm";
+export const LLM_PROVIDERS_ADMIN_URL = `${LLM_ADMIN_URL}/provider`;
 
 export const LLM_CONTEXTUAL_COST_ADMIN_URL =
   "/api/admin/llm/provider-contextual-cost";

web/src/lib/llmConfig/providers.ts

@@ -0,0 +1,69 @@
+import type { IconFunctionComponent } from "@opal/types";
+import {
+  SvgCpu,
+  SvgOpenai,
+  SvgClaude,
+  SvgOllama,
+  SvgCloud,
+  SvgAws,
+  SvgOpenrouter,
+  SvgServer,
+  SvgAzure,
+  SvgGemini,
+  SvgLitellm,
+} from "@opal/icons";
+import { LLMProviderName } from "@/interfaces/llm";
+
+const PROVIDER_ICONS: Record<string, IconFunctionComponent> = {
+  [LLMProviderName.OPENAI]: SvgOpenai,
+  [LLMProviderName.ANTHROPIC]: SvgClaude,
+  [LLMProviderName.VERTEX_AI]: SvgGemini,
+  [LLMProviderName.BEDROCK]: SvgAws,
+  [LLMProviderName.AZURE]: SvgAzure,
+  litellm: SvgLitellm,
+  [LLMProviderName.OLLAMA_CHAT]: SvgOllama,
+  [LLMProviderName.OPENROUTER]: SvgOpenrouter,
+
+  // fallback
+  [LLMProviderName.CUSTOM]: SvgServer,
+};
+
+const PROVIDER_PRODUCT_NAMES: Record<string, string> = {
+  [LLMProviderName.OPENAI]: "GPT",
+  [LLMProviderName.ANTHROPIC]: "Claude",
+  [LLMProviderName.VERTEX_AI]: "Gemini",
+  [LLMProviderName.BEDROCK]: "Amazon Bedrock",
+  [LLMProviderName.AZURE]: "Azure OpenAI",
+  litellm: "LiteLLM",
+  [LLMProviderName.OLLAMA_CHAT]: "Ollama",
+  [LLMProviderName.OPENROUTER]: "OpenRouter",
+
+  // fallback
+  [LLMProviderName.CUSTOM]: "Custom Models",
+};
+
+const PROVIDER_DISPLAY_NAMES: Record<string, string> = {
+  [LLMProviderName.OPENAI]: "OpenAI",
+  [LLMProviderName.ANTHROPIC]: "Anthropic",
+  [LLMProviderName.VERTEX_AI]: "Google Cloud Vertex AI",
+  [LLMProviderName.BEDROCK]: "AWS",
+  [LLMProviderName.AZURE]: "Microsoft Azure",
+  litellm: "LiteLLM",
+  [LLMProviderName.OLLAMA_CHAT]: "Ollama",
+  [LLMProviderName.OPENROUTER]: "OpenRouter",
+
+  // fallback
+  [LLMProviderName.CUSTOM]: "Other providers or self-hosted",
+};
+
+export function getProviderProductName(providerName: string): string {
+  return PROVIDER_PRODUCT_NAMES[providerName] ?? providerName;
+}
+
+export function getProviderDisplayName(providerName: string): string {
+  return PROVIDER_DISPLAY_NAMES[providerName] ?? providerName;
+}
+
+export function getProviderIcon(providerName: string): IconFunctionComponent {
+  return PROVIDER_ICONS[providerName] ?? SvgCpu;
+}

web/src/lib/llmConfig/svc.ts

@@ -0,0 +1,71 @@
+/**
+ * LLM action functions for mutations.
+ *
+ * These are async functions for one-off actions that don't need SWR caching.
+ *
+ * Endpoints:
+ * - /api/admin/llm/test/default - Test the default LLM provider connection
+ * - /api/admin/llm/default - Set the default LLM model
+ * - /api/admin/llm/provider/{id} - Delete an LLM provider
+ */
+
+import {
+  LLM_ADMIN_URL,
+  LLM_PROVIDERS_ADMIN_URL,
+} from "@/lib/llmConfig/constants";
+
+/**
+ * Test the default LLM provider.
+ * Returns true if the default provider is configured and working, false otherwise.
+ */
+export async function testDefaultProvider(): Promise<boolean> {
+  try {
+    const response = await fetch(`${LLM_ADMIN_URL}/test/default`, {
+      method: "POST",
+    });
+    return response?.ok || false;
+  } catch {
+    return false;
+  }
+}
+
+/**
+ * Set the default LLM model.
+ * @param providerId - The provider ID
+ * @param modelName - The model name within that provider
+ * @throws Error with the detail message from the API on failure
+ */
+export async function setDefaultLlmModel(
+  providerId: number,
+  modelName: string
+): Promise<void> {
+  const response = await fetch(`${LLM_ADMIN_URL}/default`, {
+    method: "POST",
+    headers: { "Content-Type": "application/json" },
+    body: JSON.stringify({
+      provider_id: providerId,
+      model_name: modelName,
+    }),
+  });
+
+  if (!response.ok) {
+    const errorMsg = (await response.json()).detail;
+    throw new Error(errorMsg);
+  }
+}
+
+/**
+ * Delete an LLM provider.
+ * @param providerId - The provider ID to delete
+ * @throws Error with the detail message from the API on failure
+ */
+export async function deleteLlmProvider(providerId: number): Promise<void> {
+  const response = await fetch(`${LLM_PROVIDERS_ADMIN_URL}/${providerId}`, {
+    method: "DELETE",
+  });
+
+  if (!response.ok) {
+    const errorMsg = (await response.json()).detail;
+    throw new Error(errorMsg);
+  }
+}