Sameer Kankute
8864c6937c
Merge pull request #18031 from BerriAI/litellm_anthropic_claude_skills_int
...
Add support for agent skills in chat completion
2025-12-16 21:39:11 +05:30
Sameer Kankute
fc5c5d4dbe
Merge pull request #18042 from BerriAI/litellm_embedding_default_omit
...
fix: Add none to encoding_format instead of omitting it
2025-12-16 21:35:29 +05:30
Sameer Kankute
abfa869030
Merge pull request #18049 from BerriAI/litellm_bedrock_ttl_issue
...
Remove ttl field when routing to bedrock
2025-12-16 21:34:57 +05:30
Sameer Kankute
616417eaaa
Merge pull request #18037 from BerriAI/litellm_fix_vertex_ai_headers
...
fix: skip adding beta headers for vertex ai as it is not suppported
2025-12-16 21:33:35 +05:30
Sameer Kankute
de5471e3cd
Merge pull request #18036 from BerriAI/litellm_fix_responses_headers
...
Add custom headers in responses API
2025-12-16 21:30:59 +05:30
Sameer Kankute
f23f157556
fix error: Incompatible types (expression has type dict[Any, Any]
2025-12-16 21:27:42 +05:30
Mateo Di Loreto
107ea9043a
[Feature] Download Prisma binaries at build time instead of at runtime for Security Restricted environments ( #17695 )
...
* Use config file to enable prometheus metrics
* Revert "Use config file to enable prometheus metrics"
This reverts commit 15ae36e1711791c0ac0a7aa84dcec142951717f5.
* Improve hardened stack and Prisma offline flow
* Document hardened compose usage
* Remove undesired change in fastapi-sso
* Restore dashboard lockfile
* Remove unecessary tempdirs
* Document hardened/offline Docker validation flow
2025-12-16 21:25:53 +05:30
Alexsander Hamir
8f976df651
[Refactor] litellm/init.py: lazy load default encoding from client decorator ( #18059 )
2025-12-16 07:45:24 -08:00
Alexsander Hamir
f8168f5063
[Refactor] litellm/init.py: lazy load dotprompt integration ( #18056 )
2025-12-16 07:18:11 -08:00
Alexsander Hamir
feb6924efe
[Refactor] litellm/init.py: lazy load .types.utils ( #18054 )
2025-12-16 06:42:17 -08:00
Alexsander Hamir
ea06fb70fb
refactor: lazy load bedrock types ( #18053 )
2025-12-16 06:05:01 -08:00
Alexsander Hamir
6ca812130b
[Refactor] litellm/init.py: lazy load LLMClientCache ( #18008 )
2025-12-16 05:44:06 -08:00
Sameer Kankute
6ccfb8b45e
Remove ttl fiedl when routing to bedrock
2025-12-16 16:20:19 +05:30
Krish Dholakia
e31d8a1cf6
Router order parameter documentation ( #18045 )
...
* feat: Add deployment ordering for load balancing
Co-authored-by: krrishdholakia <krrishdholakia@gmail.com>
* Refactor docs to simplify load balancing and deployment ordering explanations
Co-authored-by: krrishdholakia <krrishdholakia@gmail.com>
---------
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
2025-12-16 15:57:05 +05:30
Sameer Kankute
dbcae4aca5
fix: Add none to encoding_format instead of omitting it
2025-12-16 13:23:15 +05:30
Sameer Kankute
4c482186cd
remove is_vertex param from request
2025-12-16 12:00:37 +05:30
Sameer Kankute
7a5e49a328
fix: skip adding beta headers for vertex ai as it is not suppported
2025-12-16 11:56:46 +05:30
Sameer Kankute
85907cebb5
Add custom headers in responses API
2025-12-16 11:31:15 +05:30
Krrish Dholakia
5e78ffd5fd
fix: fix docker compose quick start
2025-12-16 11:26:38 +05:30
Krrish Dholakia
6041155de0
docs: cleanup
2025-12-16 11:21:41 +05:30
Sameer Kankute
1222d9e376
Add doc and tests for agent skils
2025-12-16 10:20:21 +05:30
Sameer Kankute
bcfc77f683
Add beta headers for claude skills
2025-12-16 10:12:36 +05:30
Sameer Kankute
9802a6a19e
Add container field in streaming response
2025-12-16 10:02:07 +05:30
yuneng-jiang
0b5317c12b
Merge pull request #18030 from BerriAI/litellm_ui_milvus
...
[Feature] UI - Milvus Vector Store
2025-12-15 20:22:38 -08:00
Krish Dholakia
23e70dd90b
Litellm dev 12 15 2025 p1 ( #18028 )
...
* fix(openai/responses/guardrail_translation): fix basemodel import
ensures /v1/responses output is correctly processed by guardrail handler
* fix: fix linting error
* fix: fix linting errors
2025-12-16 09:48:49 +05:30
yuneng-jiang
0eba879179
Add ability to add embedding model for milvus
2025-12-15 20:12:44 -08:00
Ishaan Jaff
244d83ff47
[Docs] Litellm add docs vertex ai engine ( #18027 )
...
* new provider doc
* add to sidebar
* stash docs
* docs fix
* docs vertex agent engine
2025-12-15 20:11:40 -08:00
Cesar Garcia
13c0ab8985
Cleanup PR template: remove redundant fields ( #17956 )
...
- Remove Title section (already in PR title)
- Remove screenshot requirement (CI validates tests)
Co-authored-by: Krish Dholakia <krrishdholakia@gmail.com>
2025-12-16 09:39:22 +05:30
Cesar Garcia
cd9b71093a
Improve issue labeling: add component dropdown and more provider keywords ( #17957 )
...
- Replace "ML Ops Team" dropdown with "What part of LiteLLM is this about?"
with options: SDK, Proxy, UI Dashboard, Docs, Other
- Add component dropdown to feature_request.yml template
- Rename label-mlops.yml to label-component.yml and update to auto-label
issues with sdk, proxy, ui-dashboard, or docs based on selection
- Add more provider keywords to issue-keyword-labeler: gemini, cohere,
mistral, groq, ollama, deepseek
2025-12-16 09:38:39 +05:30
Sameer Kankute
e40ad5203e
Add container field as provider specific field
2025-12-16 09:28:04 +05:30
Ishaan Jaff
ce113f4e4b
[Docs] Add docs on using pydantic ai agents with LiteLLM A2a gateway ( #18026 )
...
* init A2AProviderConfigManager
* move file
* move file
* add pydnatic ai folder
* init providers
* test_pydantic_ai_non_streaming
* fix import
* INIT pydantic
* use_a2a_form_fields
* test_vertex_agent_engine_streaming
* add agent_engine
* init transform for agent engine
* init agent engine
* VertexAgentEngineSSEStreamIterator
* sample
* ui add new fields
* fix vertex_credentials
* working SSE iterator
* TestVertexAgentEngineTransformRequest
* fix code QA check
* stash docs
* docs fix
* fix logo
* docs fix
* doc pydantic ai
* docs pydantic ai
* new provider
* docs fix
2025-12-15 19:43:12 -08:00
Krrish Dholakia
f90dea7315
fix(docker-compose.yml): move to docker.litellm.ai
2025-12-16 08:50:34 +05:30
Ishaan Jaff
32c07113cf
[Feat] New Provider - VertexAI Agent Engine ( #18014 )
...
* init A2AProviderConfigManager
* move file
* move file
* add pydnatic ai folder
* init providers
* test_pydantic_ai_non_streaming
* fix import
* INIT pydantic
* use_a2a_form_fields
* test_vertex_agent_engine_streaming
* add agent_engine
* init transform for agent engine
* init agent engine
* VertexAgentEngineSSEStreamIterator
* sample
* ui add new fields
* fix vertex_credentials
* working SSE iterator
* TestVertexAgentEngineTransformRequest
* fix code QA check
* Potential fix for code scanning alert no. 3923: Clear-text logging of sensitive information
Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
---------
Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
2025-12-15 19:14:49 -08:00
Krish Dholakia
f58b76aee8
Revert "Revert "Litellm bedrock guardrails block precedence over masking (#17…" ( #18023 )
...
This reverts commit 0abe5cdce9
2025-12-16 08:42:10 +05:30
Krish Dholakia
0abe5cdce9
Revert "Litellm bedrock guardrails block precedence over masking ( #17968 )" ( #18022 )
...
This reverts commit a0754f1c88
2025-12-16 08:41:36 +05:30
kothamah
a0754f1c88
Litellm bedrock guardrails block precedence over masking ( #17968 )
...
* prioritized bedrock guardrail blocking by removing early return based on masking flags
When mask_request_content: true or mask_response_content: true, the method immediately returning False.
The Result: Even when Bedrock Guardrails returned action: "BLOCKED" for dangerous content, LiteLLM would not raise an exception and allowing the content through the response.
So removed that early condition which will return true for the blocked actions based on guardrails.
* Added test case for bedrock guardrail block content precedence
2025-12-16 08:41:13 +05:30
Krrish Dholakia
7c2478b70e
docs: replace ghcr link with docker.litellm.ai
2025-12-16 08:35:45 +05:30
Doni Crosby
e83f08c24d
feat(venice.ai): add support for Venice.ai API via providers.json ( #17962 )
2025-12-16 08:24:16 +05:30
Dmitrii Komarov
df9d3abf49
Propagate token usage when generating images with Gemini ( #17987 )
2025-12-16 08:20:09 +05:30
Krish Dholakia
edfb4148dc
Add workflow to create daily staging branches ( #18020 )
...
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
2025-12-16 08:13:54 +05:30
yuneng-jiang
be4f11eacc
Fixing build
2025-12-15 18:38:41 -08:00
yuneng-jiang
8e3eb33165
Tests
2025-12-15 18:36:01 -08:00
yuneng-jiang
0ee924f91e
Adding svg
2025-12-15 18:34:34 -08:00
yuneng-jiang
fabbd28b45
Adding Milvus to Vector Store in UI
2025-12-15 18:31:35 -08:00
Shivam Rawat
c754794bc3
[fix] add qwen3-embedding-8b input per token price ( #18018 )
...
* added embedding input token price
* added embedding input token price
2025-12-16 07:38:21 +05:30
Krrish Dholakia
b0a9c85a52
fix: fix ruff linting errors
2025-12-16 07:38:07 +05:30
Cesar Garcia
4fdbbdfe6d
fix(anthropic): correct claude-3-7-sonnet max_tokens to 64K default ( #17979 )
...
Claude 3.7 Sonnet's default max_output_tokens is 64000, not 128000.
The 128K output limit requires the beta header 'output-128k-2025-02-19'.
This fixes the integration test failure where requests with max_tokens=128000
were being rejected by the Anthropic API.
Fixes test_multiturn_tool_calls in test_anthropic_responses_api.py
2025-12-16 07:27:40 +05:30
Ishaan Jaff
a4fb0df028
[Feat] New provider - Agent Gateway, add pydantic ai agents ( #18013 )
...
* init A2AProviderConfigManager
* move file
* move file
* add pydnatic ai folder
* init providers
* test_pydantic_ai_non_streaming
* fix import
* INIT pydantic
* use_a2a_form_fields
* TestPydanticAITransformation
2025-12-15 17:40:58 -08:00
YutaSaito
867ac17677
Merge pull request #18010 from BerriAI/litellm_fix_mcp_deepcopy_error
...
fix: mcp deepcopy error
2025-12-16 07:37:25 +09:00
Cesar Garcia
df19a747a2
feat(custom_llm): add image_edit and aimage_edit support ( #17999 )
...
* feat(custom_llm): add image_edit and aimage_edit support
Add support for image_edit and aimage_edit methods in CustomLLM class,
allowing users to implement custom image editing providers.
Changes:
- Add image_edit() and aimage_edit() methods to CustomLLM base class
- Add custom provider detection in litellm.image_edit() function
- Add tests for sync and async image_edit with custom handlers
* docs: add image_edit to CustomLLM documentation
- Add /v1/images/edits to supported routes
- Add Image Edit section with example
- Update Custom Handler Spec with image_edit methods
2025-12-15 14:16:36 -08:00
