scriptoriumadmin

scriptoriumadmin synced commits to refs/pull/37582/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:15 +00:00

6d9891d61a Merge 1d010d5f54 into 9bc9104102

9bc9104102 fix(proxy): log budget reservation notice once at config load (#40167)

1761fe236f feat(complexity_router): add declarative custom dimensions to the heuristic scorer (#40156)

7da6fe54b5 fix: skip one-shot Claude Code cache injection (#40175)

26d589cd28 Merge pull request #39234 from BerriAI/litellm_fix_agent_mcp_grants

Compare 270 commits »

scriptoriumadmin synced commits to refs/pull/37596/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:15 +00:00

fe8a1f4296 Merge ec5f945390 into 1af7a403c6

1af7a403c6 feat(mcp): start the named server's OAuth directly for a resource-scoped gateway flow (#39933)

9a9b4c4c25 feat(ui): show auto-router classification rate (#40192)

1a6aa98230 fix(spend): compare auto-router targets by deployment identity (#40206)

9dbfb060bd Merge pull request #39668 from BerriAI/litellm_lit6899_vertex_batch_tuned_endpoints

Compare 288 commits »

scriptoriumadmin synced commits to refs/pull/37597/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:15 +00:00

082769fdb7 Merge 8070294564 into 35451ecc7b

35451ecc7b fix(router): keep deployment tags out of retry and fallback tag routing (#40226)

9fffda4117 fix(router): give cost-based routing its own cache key so it stops overwriting latency samples (#40225)

3496ab9518 fix(router): let simple-shuffle weight by any deployment's weight/rpm/tpm (#40222)

935c7190eb fix(router): count allowed_fails in the shared router cache so multi-worker proxies bench a deployment fleet-wide (#40224)

Compare 257 commits »

scriptoriumadmin synced commits to refs/pull/37630/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:15 +00:00

5c7972fec7 Merge afaa5c82e694c96a96ee57a46e2586b3f8adc1b3 into ee1a6407cb

ee1a6407cb fix(guardrails): keep guardrail telemetry when a policy pipeline blocks or modifies the response (#40211)

82e6b84f5a Merge pull request #39384 from BerriAI/litellm_fix_invalid_index_after_migration_deadlock

84a5136f45 Merge remote-tracking branch 'origin/litellm_internal_staging' into litellm_fix_invalid_index_after_migration_deadlock

9b74e6f34e Merge pull request #40009 from BerriAI/litellm_lit_7039_least_busy_shared_counts

Compare 239 commits »

scriptoriumadmin synced commits to refs/pull/37296/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:14 +00:00

9c9b0741d9 Merge da8e2c4d99 into ecd8bab0a6

ecd8bab0a6 Merge pull request #40195 from BerriAI/litellm_remove_static_cost_map_tests

adcfe8cb7f test: pin redirected xai slugs to the target's tier field set

dc035cba62 test: preserve live xai pricing invariant

5cfe20a68d test: collapse blank lines left by removed tests

Compare 277 commits »

scriptoriumadmin synced commits to refs/pull/37464/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:14 +00:00

a8a91a1dec Merge 1f09b867e1 into c899912f49

1f09b867e1 chore: merge litellm_internal_staging, keep Unpack kwargs and object payload typing

c899912f49 Merge pull request #40287 from BerriAI/litellm_/release-version-bump-787548

0175c7da1c feat(router): resolve max_tokens to the tier model's ceiling on auto-routed requests (#40209)

9e18526887 feat(deploy): metrics sidecar and separate metrics port in Helm and Terraform (#40163)

Compare 972 commits »

scriptoriumadmin synced commits to refs/pull/37477/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:14 +00:00

950f996845 Merge 7c82ad24580e69227ddf11a92f4d361c5b6914ec into 1af7a403c6

1af7a403c6 feat(mcp): start the named server's OAuth directly for a resource-scoped gateway flow (#39933)

9a9b4c4c25 feat(ui): show auto-router classification rate (#40192)

1a6aa98230 fix(spend): compare auto-router targets by deployment identity (#40206)

9dbfb060bd Merge pull request #39668 from BerriAI/litellm_lit6899_vertex_batch_tuned_endpoints

Compare 217 commits »

scriptoriumadmin synced commits to refs/pull/37480/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:14 +00:00

0d6f970c75 Merge a89d5314fc3830baf559625210c62cad03d5ae7a into c5ec2eedc1

c5ec2eedc1 fix(spend): price caching savings on the billed request basis (#40160)

038025ba5e fix(guardrails): accept on_violation block and alert for mcp_security (#40155)

f896df1b06 docs: fix stale file paths in ARCHITECTURE.md (#40157)

7c1745cc71 feat(ui): make automatic auto-router setup discoverable and show what it configured (#40146)

Compare 121 commits »

scriptoriumadmin synced commits to refs/pull/37318/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:14 +00:00

8b2b22439b Merge 784e9eb3c9 into 2700ffe8c9

2700ffe8c9 fix: capture provider request id in failure logging payloads (#40045)

9275cf42ed Merge pull request #40015 from BerriAI/litellm_fix_check_run_name_collisions

8e7f748e8a Merge pull request #39870 from BerriAI/litellm_lit_6917_make_check_test_tree_ruff

c8ef043087 Merge pull request #39981 from BerriAI/litellm_lit_7079_bridge_preserve_provider_metadata

Compare 86 commits »

scriptoriumadmin synced commits to refs/pull/37313/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:14 +00:00

f001a06344 Merge cda3d8bd33 into 6a425a5cc5

6a425a5cc5 fix(responses): record spend for native Responses API WebSocket sessions (#38856)

d36e032241 fix(proxy): initialize string success/failure callbacks at startup after config load (#38226)

77bfeb5708 Merge pull request #40269 from BerriAI/litellm_lit7212_responses_bridge_keep_system_in_input

01c68c199b fix(guardrails): allow framework-supported logging-only mode (#40267)

Compare 246 commits »

scriptoriumadmin synced commits to refs/pull/37464/head at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:14 +00:00

1f09b867e1 chore: merge litellm_internal_staging, keep Unpack kwargs and object payload typing

6a425a5cc5 fix(responses): record spend for native Responses API WebSocket sessions (#38856)

d36e032241 fix(proxy): initialize string success/failure callbacks at startup after config load (#38226)

77bfeb5708 Merge pull request #40269 from BerriAI/litellm_lit7212_responses_bridge_keep_system_in_input

01c68c199b fix(guardrails): allow framework-supported logging-only mode (#40267)

Compare 3247 commits »

scriptoriumadmin synced commits to refs/pull/37460/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:14 +00:00

30b99888b9 Merge 3ec1414937 into 1af7a403c6

1af7a403c6 feat(mcp): start the named server's OAuth directly for a resource-scoped gateway flow (#39933)

9a9b4c4c25 feat(ui): show auto-router classification rate (#40192)

1a6aa98230 fix(spend): compare auto-router targets by deployment identity (#40206)

9dbfb060bd Merge pull request #39668 from BerriAI/litellm_lit6899_vertex_batch_tuned_endpoints

Compare 288 commits »

scriptoriumadmin synced commits to refs/pull/37433/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:14 +00:00

1e73ef12a0 Merge bb3a8dedb5f8dab82ea37f98e755588be10ed0c1 into 6a425a5cc5

6a425a5cc5 fix(responses): record spend for native Responses API WebSocket sessions (#38856)

d36e032241 fix(proxy): initialize string success/failure callbacks at startup after config load (#38226)

77bfeb5708 Merge pull request #40269 from BerriAI/litellm_lit7212_responses_bridge_keep_system_in_input

01c68c199b fix(guardrails): allow framework-supported logging-only mode (#40267)

Compare 245 commits »

scriptoriumadmin synced commits to refs/pull/37431/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:14 +00:00

7789f224ff Merge 7c7dbe25daf77f3ab501f4ff4567a54692e0f687 into 6a425a5cc5

6a425a5cc5 fix(responses): record spend for native Responses API WebSocket sessions (#38856)

d36e032241 fix(proxy): initialize string success/failure callbacks at startup after config load (#38226)

77bfeb5708 Merge pull request #40269 from BerriAI/litellm_lit7212_responses_bridge_keep_system_in_input

01c68c199b fix(guardrails): allow framework-supported logging-only mode (#40267)

Compare 246 commits »

scriptoriumadmin synced commits to refs/pull/37369/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:14 +00:00

e997a04506 Merge 7cd5d46c36 into 168a0055a2

168a0055a2 chore(lint): stop ratcheting *-budget.json on PR branches (#39937)

02522a5441 Merge pull request #39983 from BerriAI/litellm_lit_7081_azure_ai_gpt_6_astra_pricing

b09b7d3eb8 Merge pull request #39980 from BerriAI/litellm_lit_7048_batch_cost_row_once

defd8661f4 refactor(spend): stop queueing a batch's claim row for a writer the proxy never builds

Compare 30 commits »

scriptoriumadmin synced commits to refs/pull/37256/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:13 +00:00

152223ba58 Merge e7f9050efe into 9fffda4117

9fffda4117 fix(router): give cost-based routing its own cache key so it stops overwriting latency samples (#40225)

3496ab9518 fix(router): let simple-shuffle weight by any deployment's weight/rpm/tpm (#40222)

935c7190eb fix(router): count allowed_fails in the shared router cache so multi-worker proxies bench a deployment fleet-wide (#40224)

b0071f363f fix(router): keep per-request routing_strategy override selectors out of global callbacks (#40229)

Compare 326 commits »

scriptoriumadmin synced commits to refs/pull/37145/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:13 +00:00

edc76fce43 Merge aad8f45efc into 9b74e6f34e

9b74e6f34e Merge pull request #40009 from BerriAI/litellm_lit_7039_least_busy_shared_counts

f769aa4675 fix(router): give cooldowns their own cache so siblings see a bench in ~1s (#40025)

eeef03f122 chore: merge litellm_internal_staging into litellm_lit_7039_least_busy_shared_counts

a85c3152ca fix(proxy): load db credentials in the model reconcile so a worker never serves a model before its credential (#39876)

Compare 352 commits »

scriptoriumadmin synced commits to refs/pull/37145/head at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:13 +00:00

aad8f45efc Merge branch 'litellm_internal_staging' into fix/gemini-rolling-latest-alias-version-detection

168a0055a2 chore(lint): stop ratcheting *-budget.json on PR branches (#39937)

02522a5441 Merge pull request #39983 from BerriAI/litellm_lit_7081_azure_ai_gpt_6_astra_pricing

b09b7d3eb8 Merge pull request #39980 from BerriAI/litellm_lit_7048_batch_cost_row_once

defd8661f4 refactor(spend): stop queueing a batch's claim row for a writer the proxy never builds

Compare 4094 commits »

scriptoriumadmin synced commits to refs/pull/37254/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:13 +00:00

445f993952 Merge 05e89d9ccb into 9fffda4117

9fffda4117 fix(router): give cost-based routing its own cache key so it stops overwriting latency samples (#40225)

3496ab9518 fix(router): let simple-shuffle weight by any deployment's weight/rpm/tpm (#40222)

935c7190eb fix(router): count allowed_fails in the shared router cache so multi-worker proxies bench a deployment fleet-wide (#40224)

b0071f363f fix(router): keep per-request routing_strategy override selectors out of global callbacks (#40229)

Compare 326 commits »

scriptoriumadmin synced commits to refs/pull/37162/merge at scriptoriumadmin/litellm from mirror 2026-09-08 22:22:13 +00:00

f353edf5c5 Merge 03af235ad5 into 1af7a403c6

1af7a403c6 feat(mcp): start the named server's OAuth directly for a resource-scoped gateway flow (#39933)

9a9b4c4c25 feat(ui): show auto-router classification rate (#40192)

1a6aa98230 fix(spend): compare auto-router targets by deployment identity (#40206)

9dbfb060bd Merge pull request #39668 from BerriAI/litellm_lit6899_vertex_batch_tuned_endpoints

Compare 218 commits »