Kent 6c4fd246d3 refactor(batches): drop the alias swap now that credential resolution returns the deployment model ... get_deployment_credentials_with_provider returns credentials["model"] and prepare_data_with_credentials does data.update(credentials), so the proxy batch-create path already hands litellm.acreate_batch the deployment's real provider model. _swap_alias_for_deployment_model and Router.get_deployment_model_for_alias were doing that same swap a second time. What is left is the part that is not upstream: batch endpoints pass the caller's team_id into credential resolution, and _resolve_unblocked_deployment applies the team guard to exact deployment-id lookups too. Also regenerates the dashboard API types, which the base left stale.		2026-09-14 12:06:17 +08:00
..
litellm-dashboard	refactor(batches): drop the alias swap now that credential resolution returns the deployment model	2026-09-14 12:06:17 +08:00
Dockerfile	fix(docker): bump nginx runtime to 1.31.5-alpine3.24 and pin digest to resolve critical CVEs (#39561)	2026-09-03 08:46:07 -07:00
nginx.conf	fix(ui): boot the UI image as an arbitrary uid by anchoring nginx writes under /tmp (#37982)	2026-08-24 11:57:36 -07:00