mirror of
https://github.com/BerriAI/litellm.git
synced 2026-10-16 04:29:31 +00:00
30b8db325e
* feat(ui): show decision models as decisions and list the supported ones in the AI Hub * fix(ui): gate the decision garden's Add model links and sort the Mode column by its label The garden shows Add model only for providers the Add Model form can create and only when the session can add models, since a view-only admin's session role reads as Admin. The AI Hub's Mode column sorts by the label it shows, and the Models page parses its tab slug as a string literal * fix(ui): label the make-public dialog's decision models, clear the prefill after the add, and gate the garden's Add model by the Models page's rule * fix(ui): drop the dead decisions billing-row filter and the unread team fetch on the AI Hub The <provider>/decisions/<model> cost map keys never shipped, so the garden filter matched nothing. The AI Hub page fetched /team/list for every visitor and read it for nobody, since only proxy admins reach the add-model check. The org-admin page test now asserts the role string a session carries * fix(ui): keep self-hosted decision models off the /v1/decisions claim, order the garden one per provider, and show the decisions example on the public hub * fix(ui): mark decision models this proxy cannot serve yet in the AI Hub garden * fix(ui): use singular wording in the decision garden header when one model is left * fix(ui): list gpt-6-luna in the decision garden, send Foundry to its deployment-name docs, and record Models page tab clicks * fix(ui): say the decisions filter matched no model in decisions mode, and mock usePathname in the Models page integration test * fix(ui): stop saying deployed decision models still need adding --------- Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|---|---|---|
| .. | ||
| litellm-dashboard | ||
| Dockerfile | ||
| nginx.conf | ||