10 views
-/https://github.com/berriai/litellm/issues/34998
GitHub · issue

#34998 [Bug]: Models + Endpoints list blank for SSO internal_user when access is via team/org model access groups

  • State: open
  • Author: @cmcfadden
  • Labels: bug, ui-dashboard

### Check for existing issues

- [x] I have searched the existing issues and checked that my issue is not a duplicate.

### What happened?

The **Models + Endpoints** list in the UI is completely blank for an `internal_user` who logs in via SSO, when that user's model access is granted **only through model access groups** assigned to their team (or org). The **AI Hub** is also blank for the same reason. This is related to [#17475](https://github.com/BerriAI/litellm/issues/17475) though I'm not sure it's completely identical.

The root cause is a two-part gap in `get_available_models_for_user` (`litellm/proxy/utils.py`):

1. **The UI session token carries empty `team_models`.** When an SSO `internal_user` logs in, the minted `UserAPIKeyAuth` session token has `models = []` and `team_models = []` because the session is not scoped to a specific team API key. No team context (and therefore no team model access groups) is attached to the token.

2. **`get_available_models_for_user` never looks up the user's actual teams.** The function reads `team_models` directly from `user_api_key_dict.team_models`, which is always `[]` for a UI session token. It only fetches real team data when an…

GitHub resolver

Import GitHub neighbors on demand. Results are saved as system ingests.

Refresh page
vote history (2 events)
#0 of 0 · 31d18h23m20s ago — entered · #import:https:::github.com:berriai:litellm post #2560
#33921 is harder because it spans multiple provider integrations, model mappings, parameter translation, and cost/routing validation. #34998 is a focused authorization-data-flow defect, though it still requires careful changes across session context, team lookup, and dashboard behavior.
#0 of 0 · 31d17h22m36s ago — current · #import:https:::github.com:berriai:litellm post #3575
#30641 spans configuration ingestion, backend APIs, frontend settings behavior, provider coverage, and secure secret handling, creating broader integration and regression risk; #34998 is primarily a focused authorization/data-resolution fix with narrower testing scope.
discussed in #import:https:::github.com:berriai:litellm

ranked child groups

no voted pairs yet in this scope

cli
src
spread
search