scriptoriumadmin

scriptoriumadmin synced commits to refs/pull/50015/head at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:42 +00:00

edc99fdefd refactor(model-cost): split S3 into storage-s3 and keep catalog rules in model-catalog

4426b68dfe chore: drop tiktoken cache files committed by accident

9e8900e3a0 fix(model-cost): keep forks safe around the import refresh

a1164e3b17 fix(model-cost): keep a router's rows when a map is adopted while it registers them

fee41c0be1 refactor(model-cost): run the cost map lifecycle on shared Rust code

Compare 11 commits »

scriptoriumadmin synced commits to refs/pull/50013/merge at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:42 +00:00

3f02ca2f0d Merge 95b3bf4d6b into 429f814a04

429f814a04 fix(cost_map): fill missing Claude capability fields from the family baseline (#49815)

dc741a6e16 fix(ui): let team admins edit models on team service account keys (#49970)

42b4cfc5bf fix(nvidia_riva): send the api key exactly once by calling riva over grpc directly (#50140)

5e91bc8e86 fix(ui): send a team admin's member key edits on the first save (#50007)

Compare 61 commits »

scriptoriumadmin synced commits to refs/pull/50013/head at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:42 +00:00

95b3bf4d6b fix(proxy): set request_deadline from body timeout, not only from headers

ad776e21a0 fix(lint): fix TRY201 bare-raise and LIT012 ReadOnly on request_deadline

c0a1c9c4b4 fix(router): clamp per-attempt timeout to remaining deadline budget

3d9adc3229 fix(router): add pre-call deadline check after backoff and fix test patching

Compare 4 commits »

scriptoriumadmin synced commits to refs/pull/50012/merge at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:42 +00:00

c7e78477bc Merge e83818e96c into b4a8bc715f

b4a8bc715f test(translation): expect thinking usage fields on Bedrock responses (#50181)

728c687418 fix(mistral): a named tool_choice is dropped, the model picks any tool (#39737)

3577aa2d94 test(translation): expect thinking usage fields on Bedrock messages (#50179)

74493f39fb test(translation): expect thinking usage fields on Bedrock chat (#50175)

Compare 75 commits »

scriptoriumadmin synced commits to refs/pull/49986/head at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:41 +00:00

114090d208 fix: bill regional uplift multipliers float() accepts, bytes included

b471e57be8 Merge remote-tracking branch 'origin/main' into beluga_decrease_anys_20261102

5a392b9e5f fix(bedrock)!: keep Claude's thinking when response_format json_schema is set (#49696)

d2b2e76374 feat(proxy): add model deprecation alerts with daily requests by key (#50022)

5c60824024 test(providers): expect 503 server_is_overloaded for an upstream overload (#50154)

Compare 74 commits »

scriptoriumadmin synced commits to refs/pull/49983/merge at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:41 +00:00

fb40550acb Merge 000232cd80 into 44b85d0c5c

000232cd80 feat(proxy): add opt-in model prefix enforcement for wildcard routes

44b85d0c5c fix(cost_calculator): keep regional and speed multipliers for per-call custom prices and log the billed per-call entry (#50063)

b401ac6aaf fix(cost): bill Batch cached tokens at the Standard cache price when a row has no Batch cache price (#50062)

85fdea0fd1 fix(guardrails): apply guardrail masks to Bedrock InvokeModel Claude requests and replies (#49947)

Compare 44 commits »

scriptoriumadmin synced commits to refs/pull/49990/head at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:41 +00:00

14145a88c5 fix(bedrock): send an effort without thinking only through the bedrock provider

ce02ebfbd0 fix(bedrock): send only the effort from output_config to Converse Claude without thinking

afb8eb7dfe fix(bedrock): read an inference profile's base_model from model_info and keep shadow eval's effort with thinking

731bbed432 fix(bedrock): send an application inference profile's effort when its base_model is set

b76bbd3c92 fix(bedrock): keep an application inference profile's effort with thinking and refuse a malformed effort

Compare 6 commits »

scriptoriumadmin synced commits to refs/pull/49983/head at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:41 +00:00

000232cd80 feat(proxy): add opt-in model prefix enforcement for wildcard routes

44b85d0c5c fix(cost_calculator): keep regional and speed multipliers for per-call custom prices and log the billed per-call entry (#50063)

b401ac6aaf fix(cost): bill Batch cached tokens at the Standard cache price when a row has no Batch cache price (#50062)

85fdea0fd1 fix(guardrails): apply guardrail masks to Bedrock InvokeModel Claude requests and replies (#49947)

74783d38ea fix(github_copilot): let per-user connections work with a custom GitHub client ID (#50116)

Compare 52 commits »

scriptoriumadmin synced commits to refs/pull/49995/head at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:41 +00:00

d1b0dea353 feat(proxy): validate callback_settings definitions at boot and gate reference changes on update

c3a5e29cc5 Merge remote-tracking branch 'origin/main' into litellm_key_callback_reference

dc741a6e16 fix(ui): let team admins edit models on team service account keys (#49970)

42b4cfc5bf fix(nvidia_riva): send the api key exactly once by calling riva over grpc directly (#50140)

5e91bc8e86 fix(ui): send a team admin's member key edits on the first save (#50007)

Compare 63 commits »

scriptoriumadmin synced commits to refs/pull/49989/head at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:41 +00:00

71757ab665 Merge pull request #1 from Allenyzh/litellm_fix_daily_activity_timezone_49989

0b18ff4d3e fix(proxy): preserve daily activity date boundaries

Compare 2 commits »

scriptoriumadmin synced commits to refs/pull/49989/merge at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:41 +00:00

65dc2fa51b Merge 71757ab665 into c5da69cd13

c5da69cd13 feat(ui): show model deprecation banner on the models page (#50080)

429f814a04 fix(cost_map): fill missing Claude capability fields from the family baseline (#49815)

dc741a6e16 fix(ui): let team admins edit models on team service account keys (#49970)

42b4cfc5bf fix(nvidia_riva): send the api key exactly once by calling riva over grpc directly (#50140)

Compare 66 commits »

scriptoriumadmin synced commits to refs/pull/49995/merge at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:41 +00:00

bbb9a5d656 Merge d1b0dea353 into 5c60824024

d1b0dea353 feat(proxy): validate callback_settings definitions at boot and gate reference changes on update

5c60824024 test(providers): expect 503 server_is_overloaded for an upstream overload (#50154)

71fc204495 test(routing): expect 503 server_is_overloaded for an upstream overload (#50155)

c3a5e29cc5 Merge remote-tracking branch 'origin/main' into litellm_key_callback_reference

Compare 66 commits »

scriptoriumadmin synced commits to refs/pull/49990/merge at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:41 +00:00

78347f9f29 Merge 07744b6b44 into 80b18d165b

80b18d165b chore(proxy): use neutral example names for baggage tag mappings (#50011)

f5e740e69a chore(azure): cite ada retirement schedule (#50044)

fd4cef22ea refactor: clean up fresh tech debt in proxy auth reads, chatgpt credential forwarding, messages drop_params reads, router context signatures and spend log reads (#46643)

48bd9b01f6 feat(skills): upload local skill ZIPs from the dashboard (#50008)

Compare 12 commits »

scriptoriumadmin synced commits to refs/pull/49986/merge at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:41 +00:00

a0e1a4e209 Merge 114090d208 into c4faa55249

114090d208 fix: bill regional uplift multipliers float() accepts, bytes included

c4faa55249 refactor(ui): align Chat with dashboard styling (#50034)

796f751603 test(providers): expect 503 server_is_overloaded for an upstream overload on the responses bridge (#50166)

71fde48016 feat(alerts): show recommended model provenance and provider notices (#50162)

Compare 80 commits »

scriptoriumadmin synced commits to refs/pull/49946/head at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:40 +00:00

76bf91316e fix(proxy): keep the Codex catalog key lookup in line with the call path and cap each catalog fetch

8660323c8b fix(proxy): keep the Codex release just asked in the catalog cache so a failed fetch keeps its retry time

8eebb745bd fix(proxy): keep FedRAMP and non-api.openai.com keys home, survive catalog redirect loops, and keep fetched Codex releases cached

b50561b0bd fix(proxy): keep OpenAI's Codex catalog when a refresh fails and lay fetched catalogs over the bundled one

783b914581 test(proxy): expect OpenAI's mirrored base_instructions dropped in the Codex catalog integration test

Compare 8 commits »

scriptoriumadmin synced commits to refs/pull/49943/merge at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:40 +00:00

4c0cf16baf Merge e5dc6671a8 into 42b4cfc5bf

e5dc6671a8 feat: update EmpirioLabs models in model_prices_and_context_window.json

42b4cfc5bf fix(nvidia_riva): send the api key exactly once by calling riva over grpc directly (#50140)

5e91bc8e86 fix(ui): send a team admin's member key edits on the first save (#50007)

906fa7ed28 feat(accounting): add opt-in SpendLogs ledger (#49856)

Compare 65 commits »

scriptoriumadmin synced commits to refs/pull/49976/head at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:40 +00:00

cda0fd6ba5 Merge remote-tracking branch 'origin/main' into litellm_enable_moyai_toggle

b18188dee2 test(e2e): retire the quiet_stack lock and batch the claude_code model registration (#49984)

df78b16941 feat(openai): bill gov.api.openai.com FedRAMP endpoint with 1.1x uplift (#49951)

83be155b65 feat(cost-map): add together_ai/moonshotai/Kimi-K3.1-Pro from the Together models API (#50069)

f56f0f0809 fix(gemini): point gemini-2.5-pro and gemini-2.5-flash at the models Google reroutes them to (#50064)

Compare 32 commits »

scriptoriumadmin synced commits to refs/pull/49958/merge at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:40 +00:00

d17712a06c Merge b52e3d9e08 into f5e740e69a

f5e740e69a chore(azure): cite ada retirement schedule (#50044)

fd4cef22ea refactor: clean up fresh tech debt in proxy auth reads, chatgpt credential forwarding, messages drop_params reads, router context signatures and spend log reads (#46643)

48bd9b01f6 feat(skills): upload local skill ZIPs from the dashboard (#50008)

664241188f feat(ui): redesign the dashboard top bar and give Aria its own page (#50009)

Compare 32 commits »

scriptoriumadmin synced commits to refs/pull/49947/head at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:40 +00:00

aac224fd89 Merge main into litellm_bedrock_invoke_guardrail_messages

6b8788eb27 fix(bedrock): backfill GovCloud service-tier prices (#50121)

d29c515444 ci(unit): stop the full history clone on the db-and-spend shard, fetch the schema merge base by sha (#50120)

8fcb016dc6 test: make three flaky unit tests deterministic (#50114)

4e769d153a fix(auth): align cloud credential loading and expiry with the Python SDKs (#49589)

Compare 69 commits »

scriptoriumadmin synced commits to refs/pull/49946/merge at scriptoriumadmin/litellm from mirror 2026-11-05 07:39:40 +00:00

bfd02b776b Merge 76bf91316e into 3577aa2d94

3577aa2d94 test(translation): expect thinking usage fields on Bedrock messages (#50179)

76bf91316e fix(proxy): keep the Codex catalog key lookup in line with the call path and cap each catalog fetch

74493f39fb test(translation): expect thinking usage fields on Bedrock chat (#50175)

9752869d84 feat(router): add the Rust router core with the host as its only boundary (#49513)

Compare 102 commits »