[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-BerriAI--litellm":3,"tool-BerriAI--litellm":64},[4,17,27,35,43,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",138956,2,"2026-04-05T11:33:21",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":44,"name":45,"github_repo":46,"description_zh":47,"stars":48,"difficulty_score":23,"last_commit_at":49,"category_tags":50,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,51,52,53,15,54,26,13,55],"数据工具","视频","插件","其他","音频",{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,54],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":76,"owner_avatar_url":77,"owner_bio":78,"owner_company":79,"owner_location":79,"owner_email":80,"owner_twitter":79,"owner_website":81,"owner_url":82,"languages":83,"stars":120,"forks":121,"last_commit_at":122,"license":123,"difficulty_score":23,"env_os":124,"env_gpu":124,"env_ram":124,"env_deps":125,"category_tags":132,"github_topics":133,"view_count":146,"oss_zip_url":79,"oss_zip_packed_at":79,"status":16,"created_at":147,"updated_at":148,"faqs":149,"releases":178},665,"BerriAI\u002Flitellm","litellm","Python SDK, Proxy Server (AI Gateway) to call 100+ LLM APIs in OpenAI (or native) format, with cost tracking, guardrails, loadbalancing and logging. [Bedrock, Azure, OpenAI, VertexAI, Cohere, Anthropic, Sagemaker, HuggingFace, VLLM, NVIDIA NIM]","litellm 是一款开源的 Python SDK 和代理服务器，旨在简化大语言模型（LLM）的调用与管理。它最核心的价值在于统一了接口标准，允许开发者通过单一的 OpenAI 格式轻松访问超过 100 种不同的 LLM 服务，涵盖 OpenAI、Anthropic、Azure、AWS Bedrock 等主流平台。\n\n针对多模型接入时 API 格式各异、切换成本高的问题，litellm 屏蔽了底层差异，让集成过程变得像调用原生 OpenAI 一样简单。此外，它还提供了成本追踪、请求日志、流量负载均衡及安全护栏等高级功能，帮助团队有效管控 AI 使用风险与预算。\n\n这款工具特别适合需要快速验证不同模型效果的开发者、构建复杂 AI 应用的研究人员，以及希望搭建统一 AI 网关的企业团队。无论是通过 Python 代码直接调用，还是部署为本地代理服务器，litellm 都提供了灵活的支持，并能兼容 LangGraph 等 Agent 框架，是构建稳定、可扩展 AI 应用的理想基础设施。","\u003Ch1 align=\"center\">\n        🚅 LiteLLM\n    \u003C\u002Fh1>\n    \u003Cp align=\"center\">\n        \u003Cp align=\"center\">Call 100+ LLMs in OpenAI format. [Bedrock, Azure, OpenAI, VertexAI, Anthropic, Groq, etc.]\n        \u003C\u002Fp>\n        \u003Cp align=\"center\">\n        \u003Ca href=\"https:\u002F\u002Frender.com\u002Fdeploy?repo=https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\" target=\"_blank\" rel=\"nofollow\">\u003Cimg src=\"https:\u002F\u002Frender.com\u002Fimages\u002Fdeploy-to-render-button.svg\" alt=\"Deploy to Render\">\u003C\u002Fa>\n        \u003Ca href=\"https:\u002F\u002Frailway.app\u002Ftemplate\u002FHLP0Ub?referralCode=jch2ME\">\n          \u003Cimg src=\"https:\u002F\u002Frailway.app\u002Fbutton.svg\" alt=\"Deploy on Railway\">\n        \u003C\u002Fa>\n        \u003C\u002Fp>\n    \u003C\u002Fp>\n\u003Ch4 align=\"center\">\u003Ca href=\"https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fsimple_proxy\" target=\"_blank\">LiteLLM Proxy Server (AI Gateway)\u003C\u002Fa> | \u003Ca href=\"https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fenterprise#hosted-litellm-proxy\" target=\"_blank\"> Hosted Proxy\u003C\u002Fa> | \u003Ca href=\"https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fenterprise\"target=\"_blank\">Enterprise Tier\u003C\u002Fa>\u003C\u002Fh4>\n\u003Ch4 align=\"center\">\n    \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Flitellm\u002F\" target=\"_blank\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Flitellm.svg\" alt=\"PyPI Version\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fwww.ycombinator.com\u002Fcompanies\u002Fberriai\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FY%20Combinator-W23-orange?style=flat-square\" alt=\"Y Combinator W23\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fwa.link\u002Fhuol9n\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fstatic\u002Fv1?label=Chat%20on&message=WhatsApp&color=success&logo=WhatsApp&style=flat-square\" alt=\"Whatsapp\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FwuPM9dRgDw\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fstatic\u002Fv1?label=Chat%20on&message=Discord&color=blue&logo=Discord&style=flat-square\" alt=\"Discord\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fwww.litellm.ai\u002Fsupport\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fstatic\u002Fv1?label=Chat%20on&message=Slack&color=black&logo=Slack&style=flat-square\" alt=\"Slack\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fcodspeed.io\u002FBerriAI\u002Flitellm?utm_source=badge\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fendpoint?url=https:\u002F\u002Fcodspeed.io\u002Fbadge.json\" alt=\"CodSpeed\"\u002F>\n    \u003C\u002Fa>\n\u003C\u002Fh4>\n\n\u003Cimg width=\"2688\" height=\"1600\" alt=\"Group 7154 (1)\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_e00112788a99.png\" \u002F>\n\n\n## Use LiteLLM for\n\n\u003Cdetails open>\n\u003Csummary>\u003Cb>LLMs\u003C\u002Fb> - Call 100+ LLMs (Python SDK + AI Gateway)\u003C\u002Fsummary>\n\n[**All Supported Endpoints**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fsupported_endpoints) - `\u002Fchat\u002Fcompletions`, `\u002Fresponses`, `\u002Fembeddings`, `\u002Fimages`, `\u002Faudio`, `\u002Fbatches`, `\u002Frerank`, `\u002Fa2a`, `\u002Fmessages` and more.\n\n### Python SDK\n\n```shell\npip install litellm\n```\n\n```python\nfrom litellm import completion\nimport os\n\nos.environ[\"OPENAI_API_KEY\"] = \"your-openai-key\"\nos.environ[\"ANTHROPIC_API_KEY\"] = \"your-anthropic-key\"\n\n# OpenAI\nresponse = completion(model=\"openai\u002Fgpt-4o\", messages=[{\"role\": \"user\", \"content\": \"Hello!\"}])\n\n# Anthropic  \nresponse = completion(model=\"anthropic\u002Fclaude-sonnet-4-20250514\", messages=[{\"role\": \"user\", \"content\": \"Hello!\"}])\n```\n\n### AI Gateway (Proxy Server)\n\n[**Getting Started - E2E Tutorial**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproxy\u002Fdocker_quick_start) - Setup virtual keys, make your first request\n\n```shell\npip install 'litellm[proxy]'\nlitellm --model gpt-4o\n```\n\n```python\nimport openai\n\nclient = openai.OpenAI(api_key=\"anything\", base_url=\"http:\u002F\u002F0.0.0.0:4000\")\nresponse = client.chat.completions.create(\n    model=\"gpt-4o\",\n    messages=[{\"role\": \"user\", \"content\": \"Hello!\"}]\n)\n```\n\n[**Docs: LLM Providers**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders)\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>Agents\u003C\u002Fb> - Invoke A2A Agents (Python SDK + AI Gateway)\u003C\u002Fsummary>\n\n[**Supported Providers**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fa2a#add-a2a-agents) - LangGraph, Vertex AI Agent Engine, Azure AI Foundry, Bedrock AgentCore, Pydantic AI\n\n### Python SDK - A2A Protocol\n\n```python\nfrom litellm.a2a_protocol import A2AClient\nfrom a2a.types import SendMessageRequest, MessageSendParams\nfrom uuid import uuid4\n\nclient = A2AClient(base_url=\"http:\u002F\u002Flocalhost:10001\")\n\nrequest = SendMessageRequest(\n    id=str(uuid4()),\n    params=MessageSendParams(\n        message={\n            \"role\": \"user\",\n            \"parts\": [{\"kind\": \"text\", \"text\": \"Hello!\"}],\n            \"messageId\": uuid4().hex,\n        }\n    )\n)\nresponse = await client.send_message(request)\n```\n\n### AI Gateway (Proxy Server)\n\n**Step 1.** [Add your Agent to the AI Gateway](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fa2a#adding-your-agent)\n\n**Step 2.** Call Agent via A2A SDK\n\n```python\nfrom a2a.client import A2ACardResolver, A2AClient\nfrom a2a.types import MessageSendParams, SendMessageRequest\nfrom uuid import uuid4\nimport httpx\n\nbase_url = \"http:\u002F\u002Flocalhost:4000\u002Fa2a\u002Fmy-agent\"  # LiteLLM proxy + agent name\nheaders = {\"Authorization\": \"Bearer sk-1234\"}    # LiteLLM Virtual Key\n\nasync with httpx.AsyncClient(headers=headers) as httpx_client:\n    resolver = A2ACardResolver(httpx_client=httpx_client, base_url=base_url)\n    agent_card = await resolver.get_agent_card()\n    client = A2AClient(httpx_client=httpx_client, agent_card=agent_card)\n\n    request = SendMessageRequest(\n        id=str(uuid4()),\n        params=MessageSendParams(\n            message={\n                \"role\": \"user\",\n                \"parts\": [{\"kind\": \"text\", \"text\": \"Hello!\"}],\n                \"messageId\": uuid4().hex,\n            }\n        )\n    )\n    response = await client.send_message(request)\n```\n\n[**Docs: A2A Agent Gateway**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fa2a)\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>MCP Tools\u003C\u002Fb> - Connect MCP servers to any LLM (Python SDK + AI Gateway)\u003C\u002Fsummary>\n\n### Python SDK - MCP Bridge\n\n```python\nfrom mcp import ClientSession, StdioServerParameters\nfrom mcp.client.stdio import stdio_client\nfrom litellm import experimental_mcp_client\nimport litellm\n\nserver_params = StdioServerParameters(command=\"python\", args=[\"mcp_server.py\"])\n\nasync with stdio_client(server_params) as (read, write):\n    async with ClientSession(read, write) as session:\n        await session.initialize()\n\n        # Load MCP tools in OpenAI format\n        tools = await experimental_mcp_client.load_mcp_tools(session=session, format=\"openai\")\n\n        # Use with any LiteLLM model\n        response = await litellm.acompletion(\n            model=\"gpt-4o\",\n            messages=[{\"role\": \"user\", \"content\": \"What's 3 + 5?\"}],\n            tools=tools\n        )\n```\n\n### AI Gateway - MCP Gateway\n\n**Step 1.** [Add your MCP Server to the AI Gateway](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fmcp#adding-your-mcp)\n\n**Step 2.** Call MCP tools via `\u002Fchat\u002Fcompletions`\n\n```bash\ncurl -X POST 'http:\u002F\u002F0.0.0.0:4000\u002Fv1\u002Fchat\u002Fcompletions' \\\n  -H 'Authorization: Bearer sk-1234' \\\n  -H 'Content-Type: application\u002Fjson' \\\n  -d '{\n    \"model\": \"gpt-4o\",\n    \"messages\": [{\"role\": \"user\", \"content\": \"Summarize the latest open PR\"}],\n    \"tools\": [{\n      \"type\": \"mcp\",\n      \"server_url\": \"litellm_proxy\u002Fmcp\u002Fgithub\",\n      \"server_label\": \"github_mcp\",\n      \"require_approval\": \"never\"\n    }]\n  }'\n```\n\n### Use with Cursor IDE\n\n```json\n{\n  \"mcpServers\": {\n    \"LiteLLM\": {\n      \"url\": \"http:\u002F\u002Flocalhost:4000\u002Fmcp\u002F\",\n      \"headers\": {\n        \"x-litellm-api-key\": \"Bearer sk-1234\"\n      }\n    }\n  }\n}\n```\n\n[**Docs: MCP Gateway**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fmcp)\n\n\u003C\u002Fdetails>\n\n---\n\n## How to use LiteLLM\n\nYou can use LiteLLM through either the Proxy Server or Python SDK. Both gives you a unified interface to access multiple LLMs (100+ LLMs). Choose the option that best fits your needs:\n\n\u003Ctable style={{width: '100%', tableLayout: 'fixed'}}>\n\u003Cthead>\n\u003Ctr>\n\u003Cth style={{width: '14%'}}>\u003C\u002Fth>\n\u003Cth style={{width: '43%'}}>\u003Cstrong>\u003Ca href=\"https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fsimple_proxy\">LiteLLM AI Gateway\u003C\u002Fa>\u003C\u002Fstrong>\u003C\u002Fth>\n\u003Cth style={{width: '43%'}}>\u003Cstrong>\u003Ca href=\"https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002F\">LiteLLM Python SDK\u003C\u002Fa>\u003C\u002Fstrong>\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003C\u002Fthead>\n\u003Ctbody>\n\u003Ctr>\n\u003Ctd style={{width: '14%'}}>\u003Cstrong>Use Case\u003C\u002Fstrong>\u003C\u002Ftd>\n\u003Ctd style={{width: '43%'}}>Central service (LLM Gateway) to access multiple LLMs\u003C\u002Ftd>\n\u003Ctd style={{width: '43%'}}>Use LiteLLM directly in your Python code\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd style={{width: '14%'}}>\u003Cstrong>Who Uses It?\u003C\u002Fstrong>\u003C\u002Ftd>\n\u003Ctd style={{width: '43%'}}>Gen AI Enablement \u002F ML Platform Teams\u003C\u002Ftd>\n\u003Ctd style={{width: '43%'}}>Developers building LLM projects\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd style={{width: '14%'}}>\u003Cstrong>Key Features\u003C\u002Fstrong>\u003C\u002Ftd>\n\u003Ctd style={{width: '43%'}}>Centralized API gateway with authentication and authorization, multi-tenant cost tracking and spend management per project\u002Fuser, per-project customization (logging, guardrails, caching), virtual keys for secure access control, admin dashboard UI for monitoring and management\u003C\u002Ftd>\n\u003Ctd style={{width: '43%'}}>Direct Python library integration in your codebase, Router with retry\u002Ffallback logic across multiple deployments (e.g. Azure\u002FOpenAI) - \u003Ca href=\"https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Frouting\">Router\u003C\u002Fa>, application-level load balancing and cost tracking, exception handling with OpenAI-compatible errors, observability callbacks (Lunary, MLflow, Langfuse, etc.)\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftbody>\n\u003C\u002Ftable>\n\nLiteLLM Performance: **8ms P95 latency** at 1k RPS (See benchmarks [here](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fbenchmarks))\n\n[**Jump to LiteLLM Proxy (LLM Gateway) Docs**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fsimple_proxy) \u003Cbr>\n[**Jump to Supported LLM Providers**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders)\n\n**Stable Release:** Use docker images with the `-stable` tag. These have undergone 12 hour load tests, before being published. [More information about the release cycle here](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproxy\u002Frelease_cycle)\n\nSupport for more providers. Missing a provider or LLM Platform, raise a [feature request](https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fissues\u002Fnew?assignees=&labels=enhancement&projects=&template=feature_request.yml&title=%5BFeature%5D%3A+).\n\n## OSS Adopters \n\n\u003Ctable>\n  \u003Ctr>\n    \u003Ctd>\u003Cimg height=\"60\" alt=\"Stripe\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_546bc44c798f.png\" \u002F>\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg height=\"60\" alt=\"image\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_6fbedff6d885.png\" \u002F>\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg height=\"60\" alt=\"Google ADK\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_5d7bc8f4f928.png\" \u002F>\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg height=\"60\" alt=\"Greptile\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_b06e3f141861.png\" \u002F>\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg height=\"60\" alt=\"OpenHands\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_965548da6865.png\" \u002F>\u003C\u002Ftd>\n    \u003Ctd>\u003Ch2>Netflix\u003C\u002Fh2>\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg height=\"60\" alt=\"OpenAI Agents SDK\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_a649373c69b7.png\" \u002F>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n\u003C\u002Ftable>\n\n## Supported Providers ([Website Supported Models](https:\u002F\u002Fmodels.litellm.ai\u002F) | [Docs](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders))\n\n| Provider                                                                            | `\u002Fchat\u002Fcompletions` | `\u002Fmessages` | `\u002Fresponses` | `\u002Fembeddings` | `\u002Fimage\u002Fgenerations` | `\u002Faudio\u002Ftranscriptions` | `\u002Faudio\u002Fspeech` | `\u002Fmoderations` | `\u002Fbatches` | `\u002Frerank` |\n|-------------------------------------------------------------------------------------|---------------------|-------------|--------------|---------------|----------------------|-------------------------|-----------------|----------------|-----------|-----------|\n| [Abliteration (`abliteration`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fabliteration) | ✅ |  |  |  |  |  |  |  |  |  |\n| [AI\u002FML API (`aiml`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Faiml) | ✅ | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |\n| [AI21 (`ai21`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fai21) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [AI21 Chat (`ai21_chat`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fai21) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Aleph Alpha](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Faleph_alpha) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Amazon Nova](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Famazon_nova) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Anthropic (`anthropic`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fanthropic) | ✅ | ✅ | ✅ |  |  |  |  |  | ✅ |  |\n| [Anthropic Text (`anthropic_text`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fanthropic) | ✅ | ✅ | ✅ |  |  |  |  |  | ✅ |  |\n| [Anyscale](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fanyscale) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [AssemblyAI (`assemblyai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fpass_through\u002Fassembly_ai) | ✅ | ✅ | ✅ |  |  | ✅ |  |  |  |  |\n| [Auto Router (`auto_router`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproxy\u002Fauto_routing) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [AWS - Bedrock (`bedrock`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fbedrock) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  | ✅ |\n| [AWS - Sagemaker (`sagemaker`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Faws_sagemaker) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [Azure (`azure`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fazure) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |  |\n| [Azure AI (`azure_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fazure_ai) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |  |\n| [Azure Text (`azure_text`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fazure) | ✅ | ✅ | ✅ |  |  | ✅ | ✅ | ✅ | ✅ |  |\n| [Baseten (`baseten`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fbaseten) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Bytez (`bytez`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fbytez) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Cerebras (`cerebras`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcerebras) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Clarifai (`clarifai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fclarifai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Cloudflare AI Workers (`cloudflare`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcloudflare_workers) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Codestral (`codestral`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcodestral) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Cohere (`cohere`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcohere) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  | ✅ |\n| [Cohere Chat (`cohere_chat`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcohere) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [CometAPI (`cometapi`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcometapi) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [CompactifAI (`compactifai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcompactifai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Custom (`custom`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcustom_llm_server) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Custom OpenAI (`custom_openai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fopenai_compatible) | ✅ | ✅ | ✅ |  |  | ✅ | ✅ | ✅ | ✅ |  |\n| [Dashscope (`dashscope`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fdashscope) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Databricks (`databricks`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fdatabricks) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [DataRobot (`datarobot`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fdatarobot) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Deepgram (`deepgram`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fdeepgram) | ✅ | ✅ | ✅ |  |  | ✅ |  |  |  |  |\n| [DeepInfra (`deepinfra`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fdeepinfra) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Deepseek (`deepseek`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fdeepseek) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [ElevenLabs (`elevenlabs`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Felevenlabs) | ✅ | ✅ | ✅ |  |  | ✅ | ✅ |  |  |  |\n| [Empower (`empower`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fempower) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Fal AI (`fal_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ffal_ai) | ✅ | ✅ | ✅ |  | ✅ |  |  |  |  |  |\n| [Featherless AI (`featherless_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ffeatherless_ai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Fireworks AI (`fireworks_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ffireworks_ai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [FriendliAI (`friendliai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ffriendliai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Galadriel (`galadriel`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fgaladriel) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [GitHub Copilot (`github_copilot`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fgithub_copilot) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [GitHub Models (`github`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fgithub) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Google - PaLM](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fpalm) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Google - Vertex AI (`vertex_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fvertex) | ✅ | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |\n| [Google AI Studio - Gemini (`gemini`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fgemini) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [GradientAI (`gradient_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fgradient_ai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Groq AI (`groq`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fgroq) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Heroku (`heroku`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fheroku) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Hosted VLLM (`hosted_vllm`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fvllm) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Huggingface (`huggingface`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fhuggingface) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  | ✅ |\n| [Hyperbolic (`hyperbolic`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fhyperbolic) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [IBM - Watsonx.ai (`watsonx`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fwatsonx) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [Infinity (`infinity`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Finfinity) |  |  |  | ✅ |  |  |  |  |  |  |\n| [Jina AI (`jina_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fjina_ai) |  |  |  | ✅ |  |  |  |  |  |  |\n| [Lambda AI (`lambda_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Flambda_ai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Lemonade (`lemonade`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Flemonade) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [LiteLLM Proxy (`litellm_proxy`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Flitellm_proxy) | ✅ | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |\n| [Llamafile (`llamafile`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fllamafile) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [LM Studio (`lm_studio`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Flm_studio) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Maritalk (`maritalk`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fmaritalk) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Meta - Llama API (`meta_llama`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fmeta_llama) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Mistral AI API (`mistral`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fmistral) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [Moonshot (`moonshot`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fmoonshot) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Morph (`morph`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fmorph) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Nebius AI Studio (`nebius`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fnebius) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [NLP Cloud (`nlp_cloud`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fnlp_cloud) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Novita AI (`novita`)](https:\u002F\u002Fnovita.ai\u002Fmodels\u002Fllm?utm_source=github_litellm&utm_medium=github_readme&utm_campaign=github_link) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Nscale (`nscale`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fnscale) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Nvidia NIM (`nvidia_nim`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fnvidia_nim) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [OCI (`oci`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Foci) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Ollama (`ollama`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Follama) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [Ollama Chat (`ollama_chat`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Follama) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Oobabooga (`oobabooga`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fopenai_compatible) | ✅ | ✅ | ✅ |  |  | ✅ | ✅ | ✅ | ✅ |  |\n| [OpenAI (`openai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fopenai) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |  |\n| [OpenAI-like (`openai_like`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fopenai_compatible) |  |  |  | ✅ |  |  |  |  |  |  |\n| [OpenRouter (`openrouter`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fopenrouter) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [OVHCloud AI Endpoints (`ovhcloud`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fovhcloud) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Perplexity AI (`perplexity`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fperplexity) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Petals (`petals`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fpetals) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Predibase (`predibase`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fpredibase) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Recraft (`recraft`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Frecraft) |  |  |  |  | ✅ |  |  |  |  |  |\n| [Replicate (`replicate`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Freplicate) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Sagemaker Chat (`sagemaker_chat`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Faws_sagemaker) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Sambanova (`sambanova`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fsambanova) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Snowflake (`snowflake`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fsnowflake) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Text Completion Codestral (`text-completion-codestral`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcodestral) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Text Completion OpenAI (`text-completion-openai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ftext_completion_openai) | ✅ | ✅ | ✅ |  |  | ✅ | ✅ | ✅ | ✅ |  |\n| [Together AI (`together_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ftogetherai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Topaz (`topaz`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ftopaz) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Triton (`triton`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ftriton-inference-server) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [V0 (`v0`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fv0) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Vercel AI Gateway (`vercel_ai_gateway`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fvercel_ai_gateway) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [VLLM (`vllm`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fvllm) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Volcengine (`volcengine`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fvolcano) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Voyage AI (`voyage`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fvoyage) |  |  |  | ✅ |  |  |  |  |  |  |\n| [WandB Inference (`wandb`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fwandb_inference) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Watsonx Text (`watsonx_text`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fwatsonx) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [xAI (`xai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fxai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Xinference (`xinference`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fxinference) |  |  |  | ✅ |  |  |  |  |  |  |\n\n[**Read the Docs**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002F)\n\n## Run in Developer mode\n### Services\n1. Setup .env file in root\n2. Run dependant services `docker-compose up db prometheus`\n\n### Backend\n1. (In root) create virtual environment `python -m venv .venv`\n2. Activate virtual environment `source .venv\u002Fbin\u002Factivate`\n3. Install dependencies `pip install -e \".[all]\"`\n4. `pip install prisma`\n5. `prisma generate`\n6. Start proxy backend `python litellm\u002Fproxy\u002Fproxy_cli.py`\n\n### Frontend\n1. Navigate to `ui\u002Flitellm-dashboard`\n2. Install dependencies `npm install`\n3. Run `npm run dev` to start the dashboard\n\n# Enterprise\nFor companies that need better security, user management and professional support\n\n[Talk to founders](https:\u002F\u002Fenterprise.litellm.ai\u002Fdemo)\n\nThis covers:\n- ✅ **Features under the [LiteLLM Commercial License](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproxy\u002Fenterprise):**\n- ✅ **Feature Prioritization**\n- ✅ **Custom Integrations**\n- ✅ **Professional Support - Dedicated discord + slack**\n- ✅ **Custom SLAs**\n- ✅ **Secure access with Single Sign-On**\n\n# Contributing\n\nWe welcome contributions to LiteLLM! Whether you're fixing bugs, adding features, or improving documentation, we appreciate your help.\n\n## Quick Start for Contributors\n\nThis requires poetry to be installed.\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm.git\ncd litellm\nmake install-dev    # Install development dependencies\nmake format         # Format your code\nmake lint           # Run all linting checks\nmake test-unit      # Run unit tests\nmake format-check   # Check formatting only\n```\n\nFor detailed contributing guidelines, see [CONTRIBUTING.md](CONTRIBUTING.md).\n\n## Code Quality \u002F Linting\n\nLiteLLM follows the [Google Python Style Guide](https:\u002F\u002Fgoogle.github.io\u002Fstyleguide\u002Fpyguide.html).\n\nOur automated checks include:\n- **Black** for code formatting\n- **Ruff** for linting and code quality\n- **MyPy** for type checking\n- **Circular import detection**\n- **Import safety checks**\n\n\nAll these checks must pass before your PR can be merged.\n\n\n# Support \u002F talk with founders\n\n- [Schedule Demo 👋](https:\u002F\u002Fcalendly.com\u002Fd\u002F4mp-gd3-k5k\u002Fberriai-1-1-onboarding-litellm-hosted-version)\n- [Community Discord 💭](https:\u002F\u002Fdiscord.gg\u002FwuPM9dRgDw)\n- [Community Slack 💭](https:\u002F\u002Fwww.litellm.ai\u002Fsupport)\n- Our emails ✉️ ishaan@berri.ai \u002F krrish@berri.ai\n\n# Why did we build this\n\n- **Need for simplicity**: Our code started to get extremely complicated managing & translating calls between Azure, OpenAI and Cohere.\n\n# Contributors\n\n\u003C!-- ALL-CONTRIBUTORS-LIST:START - Do not remove or modify this section -->\n\u003C!-- prettier-ignore-start -->\n\u003C!-- markdownlint-disable -->\n\n\u003C!-- markdownlint-restore -->\n\u003C!-- prettier-ignore-end -->\n\n\u003C!-- ALL-CONTRIBUTORS-LIST:END -->\n\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fgraphs\u002Fcontributors\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_a2f9195dab1c.png\" \u002F>\n\u003C\u002Fa>\n\n","\u003Ch1 align=\"center\">\n        🚅 LiteLLM\n    \u003C\u002Fh1>\n    \u003Cp align=\"center\">\n        \u003Cp align=\"center\">以 OpenAI 格式调用 100+ 个大型语言模型 (LLM)。[Bedrock, Azure, OpenAI, VertexAI, Anthropic, Groq 等]\n        \u003C\u002Fp>\n        \u003Cp align=\"center\">\n        \u003Ca href=\"https:\u002F\u002Frender.com\u002Fdeploy?repo=https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\" target=\"_blank\" rel=\"nofollow\">\u003Cimg src=\"https:\u002F\u002Frender.com\u002Fimages\u002Fdeploy-to-render-button.svg\" alt=\"部署到 Render\">\u003C\u002Fa>\n        \u003Ca href=\"https:\u002F\u002Frailway.app\u002Ftemplate\u002FHLP0Ub?referralCode=jch2ME\">\n          \u003Cimg src=\"https:\u002F\u002Frailway.app\u002Fbutton.svg\" alt=\"在 Railway 上部署\">\n        \u003C\u002Fa>\n        \u003C\u002Fp>\n    \u003C\u002Fp>\n\u003Ch4 align=\"center\">\u003Ca href=\"https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fsimple_proxy\" target=\"_blank\">LiteLLM 代理服务器 (AI 网关)\u003C\u002Fa> | \u003Ca href=\"https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fenterprise#hosted-litellm-proxy\" target=\"_blank\">托管代理\u003C\u002Fa> | \u003Ca href=\"https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fenterprise\"target=\"_blank\">企业版\u003C\u002Fa>\u003C\u002Fh4>\n\u003Ch4 align=\"center\">\n    \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Flitellm\u002F\" target=\"_blank\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Flitellm.svg\" alt=\"PyPI 版本\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fwww.ycombinator.com\u002Fcompanies\u002Fberriai\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FY%20Combinator-W23-orange?style=flat-square\" alt=\"Y Combinator W23\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fwa.link\u002Fhuol9n\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fstatic\u002Fv1?label=Chat%20on&message=WhatsApp&color=success&logo=WhatsApp&style=flat-square\" alt=\"WhatsApp\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FwuPM9dRgDw\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fstatic\u002Fv1?label=Chat%20on&message=Discord&color=blue&logo=Discord&style=flat-square\" alt=\"Discord\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fwww.litellm.ai\u002Fsupport\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fstatic\u002Fv1?label=Chat%20on&message=Slack&color=black&logo=Slack&style=flat-square\" alt=\"Slack\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fcodspeed.io\u002FBerriAI\u002Flitellm?utm_source=badge\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fendpoint?url=https:\u002F\u002Fcodspeed.io\u002Fbadge.json\" alt=\"CodSpeed\"\u002F>\n    \u003C\u002Fa>\n\u003C\u002Fh4>\n\n\u003Cimg width=\"2688\" height=\"1600\" alt=\"Group 7154 (1)\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_e00112788a99.png\" \u002F>\n\n\n## 使用 LiteLLM 进行\n\n\u003Cdetails open>\n\u003Csummary>\u003Cb>大型语言模型 (LLMs)\u003C\u002Fb> - 调用 100+ 个大型语言模型 (Python SDK + AI 网关)\u003C\u002Fsummary>\n\n[**所有支持的端点**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fsupported_endpoints) - `\u002Fchat\u002Fcompletions`, `\u002Fresponses`, `\u002Fembeddings`, `\u002Fimages`, `\u002Faudio`, `\u002Fbatches`, `\u002Frerank`, `\u002Fa2a`, `\u002Fmessages` 以及更多。\n\n### Python SDK\n\n```shell\npip install litellm\n```\n\n```python\nfrom litellm import completion\nimport os\n\nos.environ[\"OPENAI_API_KEY\"] = \"your-openai-key\"\nos.environ[\"ANTHROPIC_API_KEY\"] = \"your-anthropic-key\"\n\n# OpenAI\nresponse = completion(model=\"openai\u002Fgpt-4o\", messages=[{\"role\": \"user\", \"content\": \"Hello!\"}])\n\n# Anthropic  \nresponse = completion(model=\"anthropic\u002Fclaude-sonnet-4-20250514\", messages=[{\"role\": \"user\", \"content\": \"Hello!\"}])\n```\n\n### AI 网关 (代理服务器)\n\n[**入门指南 - 端到端教程**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproxy\u002Fdocker_quick_start) - 设置虚拟密钥，发起您的第一个请求\n\n```shell\npip install 'litellm[proxy]'\nlitellm --model gpt-4o\n```\n\n```python\nimport openai\n\nclient = openai.OpenAI(api_key=\"anything\", base_url=\"http:\u002F\u002F0.0.0.0:4000\")\nresponse = client.chat.completions.create(\n    model=\"gpt-4o\",\n    messages=[{\"role\": \"user\", \"content\": \"Hello!\"}]\n)\n```\n\n[**文档：LLM 提供商**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders)\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>智能体 (Agents)\u003C\u002Fb> - 调用 A2A 智能体 (Python SDK + AI 网关)\u003C\u002Fsummary>\n\n[**支持的提供商**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fa2a#add-a2a-agents) - LangGraph, Vertex AI Agent Engine, Azure AI Foundry, Bedrock AgentCore, Pydantic AI\n\n### Python SDK - A2A 协议\n\n```python\nfrom litellm.a2a_protocol import A2AClient\nfrom a2a.types import SendMessageRequest, MessageSendParams\nfrom uuid import uuid4\n\nclient = A2AClient(base_url=\"http:\u002F\u002Flocalhost:10001\")\n\nrequest = SendMessageRequest(\n    id=str(uuid4()),\n    params=MessageSendParams(\n        message={\n            \"role\": \"user\",\n            \"parts\": [{\"kind\": \"text\", \"text\": \"Hello!\"}],\n            \"messageId\": uuid4().hex,\n        }\n    )\n)\nresponse = await client.send_message(request)\n```\n\n### AI 网关 (代理服务器)\n\n**步骤 1.** [将您的智能体添加到 AI 网关](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fa2a#adding-your-agent)\n\n**步骤 2.** 通过 A2A SDK 调用智能体\n\n```python\nfrom a2a.client import A2ACardResolver, A2AClient\nfrom a2a.types import MessageSendParams, SendMessageRequest\nfrom uuid import uuid4\nimport httpx\n\nbase_url = \"http:\u002F\u002Flocalhost:4000\u002Fa2a\u002Fmy-agent\"  # LiteLLM 代理 + 智能体名称\nheaders = {\"Authorization\": \"Bearer sk-1234\"}    # LiteLLM 虚拟密钥\n\nasync with httpx.AsyncClient(headers=headers) as httpx_client:\n    resolver = A2ACardResolver(httpx_client=httpx_client, base_url=base_url)\n    agent_card = await resolver.get_agent_card()\n    client = A2AClient(httpx_client=httpx_client, agent_card=agent_card)\n\n    request = SendMessageRequest(\n        id=str(uuid4()),\n        params=MessageSendParams(\n            message={\n                \"role\": \"user\",\n                \"parts\": [{\"kind\": \"text\", \"text\": \"Hello!\"}],\n                \"messageId\": uuid4().hex,\n            }\n        )\n    )\n    response = await client.send_message(request)\n```\n\n[**文档：A2A 智能体网关**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fa2a)\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>MCP 工具\u003C\u002Fb> - 将 MCP 服务器连接到任意 LLM (Python SDK + AI 网关)\u003C\u002Fsummary>\n\n### Python SDK - MCP 桥接\n\n```python\nfrom mcp import ClientSession, StdioServerParameters\nfrom mcp.client.stdio import stdio_client\nfrom litellm import experimental_mcp_client\nimport litellm\n\nserver_params = StdioServerParameters(command=\"python\", args=[\"mcp_server.py\"])\n\nasync with stdio_client(server_params) as (read, write):\n    async with ClientSession(read, write) as session:\n        await session.initialize()\n\n        # Load MCP tools in OpenAI format\n        tools = await experimental_mcp_client.load_mcp_tools(session=session, format=\"openai\")\n\n        # Use with any LiteLLM model\n        response = await litellm.acompletion(\n            model=\"gpt-4o\",\n            messages=[{\"role\": \"user\", \"content\": \"What's 3 + 5?\"}],\n            tools=tools\n        )\n```\n\n### AI 网关 - MCP 网关\n\n**步骤 1.** [将您的 MCP 服务器添加到 AI 网关](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fmcp#adding-your-mcp)\n\n**步骤 2.** 通过 `\u002Fchat\u002Fcompletions` 调用 MCP 工具\n\n```bash\ncurl -X POST 'http:\u002F\u002F0.0.0.0:4000\u002Fv1\u002Fchat\u002Fcompletions' \\\n  -H 'Authorization: Bearer sk-1234' \\\n  -H 'Content-Type: application\u002Fjson' \\\n  -d '{\n    \"model\": \"gpt-4o\",\n    \"messages\": [{\"role\": \"user\", \"content\": \"Summarize the latest open PR\"}],\n    \"tools\": [{\n      \"type\": \"mcp\",\n      \"server_url\": \"litellm_proxy\u002Fmcp\u002Fgithub\",\n      \"server_label\": \"github_mcp\",\n      \"require_approval\": \"never\"\n    }]\n  }'\n```\n\n### 与 Cursor IDE 配合使用\n\n```json\n{\n  \"mcpServers\": {\n    \"LiteLLM\": {\n      \"url\": \"http:\u002F\u002Flocalhost:4000\u002Fmcp\u002F\",\n      \"headers\": {\n        \"x-litellm-api-key\": \"Bearer sk-1234\"\n      }\n    }\n  }\n}\n```\n\n[**文档：MCP 网关**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fmcp)\n\n\u003C\u002Fdetails>\n\n---\n\n## 如何使用 LiteLLM\n\n你可以通过 **代理服务器 (Proxy Server)** 或 **Python SDK** 来使用 LiteLLM。两者都为你提供统一的接口来访问多个 **大型语言模型 (LLMs)**（超过 100 个）。请选择最适合你需求的选项：\n\n\u003Ctable style={{width: '100%', tableLayout: 'fixed'}}>\n\u003Cthead>\n\u003Ctr>\n\u003Cth style={{width: '14%'}}>\u003C\u002Fth>\n\u003Cth style={{width: '43%'}}>\u003Cstrong>\u003Ca href=\"https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fsimple_proxy\">LiteLLM AI 网关\u003C\u002Fa>\u003C\u002Fstrong>\u003C\u002Fth>\n\u003Cth style={{width: '43%'}}>\u003Cstrong>\u003Ca href=\"https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002F\">LiteLLM Python SDK\u003C\u002Fa>\u003C\u002Fstrong>\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003C\u002Fthead>\n\u003Ctbody>\n\u003Ctr>\n\u003Ctd style={{width: '14%'}}>\u003Cstrong>使用场景\u003C\u002Fstrong>\u003C\u002Ftd>\n\u003Ctd style={{width: '43%'}}>中央服务（LLM 网关）以访问多个大型语言模型\u003C\u002Ftd>\n\u003Ctd style={{width: '43%'}}>在 Python 代码中直接使用 LiteLLM\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd style={{width: '14%'}}>\u003Cstrong>谁在使用？\u003C\u002Fstrong>\u003C\u002Ftd>\n\u003Ctd style={{width: '43%'}}>生成式 AI 赋能团队 \u002F 机器学习平台团队\u003C\u002Ftd>\n\u003Ctd style={{width: '43%'}}>构建 LLM 项目的开发人员\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd style={{width: '14%'}}>\u003Cstrong>关键功能\u003C\u002Fstrong>\u003C\u002Ftd>\n\u003Ctd style={{width: '43%'}}>集中式 API 网关，具备身份验证和授权功能，支持多租户成本追踪及按项目\u002F用户管理支出，支持按项目自定义（日志记录、护栏、缓存），用于安全访问控制的虚拟密钥，以及用于监控和管理的管理员仪表板 UI\u003C\u002Ftd>\n\u003Ctd style={{width: '43%'}}>直接在代码库中集成 Python 库，跨多个部署（例如 Azure\u002FOpenAI）的路由器，具备重试\u002F回退逻辑 - \u003Ca href=\"https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Frouting\">路由器\u003C\u002Fa>，应用级负载均衡和成本追踪，带有 OpenAI 兼容错误的异常处理，可观测性回调（Lunary, MLflow, Langfuse 等）\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftbody>\n\u003C\u002Ftable>\n\nLiteLLM 性能：在 1k RPS (每秒请求数) 下具有 **8ms P95 延迟**（在此查看基准测试 [here](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fbenchmarks)）\n\n[**跳转到 LiteLLM 代理（LLM 网关）文档**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fsimple_proxy) \u003Cbr>\n[**跳转到支持的 LLM 提供商**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders)\n\n**稳定版本：** 使用带有 `-stable` 标签的 Docker 镜像。这些镜像在发布前经过了 12 小时的负载测试。[关于发布周期的更多信息请见此处](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproxy\u002Frelease_cycle)\n\n支持更多提供商。如果缺少某个提供商或 LLM 平台，请提交 [功能请求](https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fissues\u002Fnew?assignees=&labels=enhancement&projects=&template=feature_request.yml&title=%5BFeature%5D%3A+)。\n\n## 开源采用者 \n\n\u003Ctable>\n  \u003Ctr>\n    \u003Ctd>\u003Cimg height=\"60\" alt=\"Stripe\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_546bc44c798f.png\" \u002F>\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg height=\"60\" alt=\"image\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_6fbedff6d885.png\" \u002F>\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg height=\"60\" alt=\"Google ADK\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_5d7bc8f4f928.png\" \u002F>\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg height=\"60\" alt=\"Greptile\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_b06e3f141861.png\" \u002F>\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg height=\"60\" alt=\"OpenHands\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_965548da6865.png\" \u002F>\u003C\u002Ftd>\n    \u003Ctd>\u003Ch2>Netflix\u003C\u002Fh2>\u003C\u002Ftd>\n    \u003Ctd>\u003Cimg height=\"60\" alt=\"OpenAI Agents SDK\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_a649373c69b7.png\" \u002F>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n\u003C\u002Ftable>\n\n## 支持的提供商（[网站支持模型](https:\u002F\u002Fmodels.litellm.ai\u002F) | [文档](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders)）\n\n| 模型提供商                                                                            | `\u002Fchat\u002Fcompletions` | `\u002Fmessages` | `\u002Fresponses` | `\u002Fembeddings` | `\u002Fimage\u002Fgenerations` | `\u002Faudio\u002Ftranscriptions` | `\u002Faudio\u002Fspeech` | `\u002Fmoderations` | `\u002Fbatches` | `\u002Frerank` |\n|-------------------------------------------------------------------------------------|---------------------|-------------|--------------|---------------|----------------------|-------------------------|-----------------|----------------|-----------|-----------|\n| [Abliteration (`abliteration`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fabliteration) | ✅ |  |  |  |  |  |  |  |  |  |\n| [AI\u002FML API (`aiml`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Faiml) | ✅ | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |\n| [AI21 (`ai21`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fai21) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [AI21 Chat (`ai21_chat`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fai21) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Aleph Alpha](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Faleph_alpha) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Amazon Nova](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Famazon_nova) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Anthropic (`anthropic`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fanthropic) | ✅ | ✅ | ✅ |  |  |  |  |  | ✅ |  |\n| [Anthropic Text (`anthropic_text`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fanthropic) | ✅ | ✅ | ✅ |  |  |  |  |  | ✅ |  |\n| [Anyscale](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fanyscale) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [AssemblyAI (`assemblyai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fpass_through\u002Fassembly_ai) | ✅ | ✅ | ✅ |  |  | ✅ |  |  |  |  |\n| [Auto Router (`auto_router`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproxy\u002Fauto_routing) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [AWS - Bedrock (`bedrock`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fbedrock) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  | ✅ |\n| [AWS - Sagemaker (`sagemaker`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Faws_sagemaker) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [Azure (`azure`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fazure) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |  |\n| [Azure AI (`azure_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fazure_ai) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |  |\n| [Azure Text (`azure_text`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fazure) | ✅ | ✅ | ✅ |  |  | ✅ | ✅ | ✅ | ✅ |  |\n| [Baseten (`baseten`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fbaseten) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Bytez (`bytez`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fbytez) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Cerebras (`cerebras`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcerebras) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Clarifai (`clarifai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fclarifai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Cloudflare AI Workers (`cloudflare`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcloudflare_workers) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Codestral (`codestral`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcodestral) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Cohere (`cohere`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcohere) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  | ✅ |\n| [Cohere Chat (`cohere_chat`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcohere) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [CometAPI (`cometapi`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcometapi) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [CompactifAI (`compactifai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcompactifai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Custom (`custom`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcustom_llm_server) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Custom OpenAI (`custom_openai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fopenai_compatible) | ✅ | ✅ | ✅ |  |  | ✅ | ✅ | ✅ | ✅ |  |\n| [Dashscope (`dashscope`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fdashscope) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Databricks (`databricks`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fdatabricks) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [DataRobot (`datarobot`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fdatarobot) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Deepgram (`deepgram`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fdeepgram) | ✅ | ✅ | ✅ |  |  | ✅ |  |  |  |  |\n| [DeepInfra (`deepinfra`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fdeepinfra) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Deepseek (`deepseek`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fdeepseek) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [ElevenLabs (`elevenlabs`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Felevenlabs) | ✅ | ✅ | ✅ |  |  | ✅ | ✅ |  |  |  |\n| [Empower (`empower`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fempower) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Fal AI (`fal_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ffal_ai) | ✅ | ✅ | ✅ |  | ✅ |  |  |  |  |  |\n| [Featherless AI (`featherless_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ffeatherless_ai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Fireworks AI (`fireworks_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ffireworks_ai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [FriendliAI (`friendliai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ffriendliai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Galadriel (`galadriel`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fgaladriel) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [GitHub Copilot (`github_copilot`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fgithub_copilot) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [GitHub Models (`github`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fgithub) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Google - PaLM](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fpalm) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Google - Vertex AI (`vertex_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fvertex) | ✅ | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |\n| [Google AI Studio - Gemini (`gemini`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fgemini) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [GradientAI (`gradient_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fgradient_ai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Groq AI (`groq`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fgroq) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Heroku (`heroku`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fheroku) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Hosted VLLM (`hosted_vllm`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fvllm) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Huggingface (`huggingface`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fhuggingface) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  | ✅ |\n| [Hyperbolic (`hyperbolic`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fhyperbolic) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [IBM - Watsonx.ai (`watsonx`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fwatsonx) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [Infinity (`infinity`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Finfinity) |  |  |  | ✅ |  |  |  |  |  |  |\n| [Jina AI (`jina_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fjina_ai) |  |  |  | ✅ |  |  |  |  |  |  |\n| [Lambda AI (`lambda_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Flambda_ai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Lemonade (`lemonade`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Flemonade) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [LiteLLM Proxy (`litellm_proxy`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Flitellm_proxy) | ✅ | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |\n| [Llamafile (`llamafile`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fllamafile) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [LM Studio (`lm_studio`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Flm_studio) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Maritalk (`maritalk`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fmaritalk) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Meta - Llama API (`meta_llama`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fmeta_llama) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Mistral AI API (`mistral`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fmistral) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [Moonshot (`moonshot`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fmoonshot) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Morph (`morph`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fmorph) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Nebius AI Studio (`nebius`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fnebius) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [NLP Cloud (`nlp_cloud`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fnlp_cloud) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Novita AI (`novita`)](https:\u002F\u002Fnovita.ai\u002Fmodels\u002Fllm?utm_source=github_litellm&utm_medium=github_readme&utm_campaign=github_link) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Nscale (`nscale`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fnscale) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Nvidia NIM (`nvidia_nim`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fnvidia_nim) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [OCI (`oci`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Foci) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Ollama (`ollama`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Follama) | ✅ | ✅ | ✅ | ✅ |  |  |  |  |  |  |\n| [Ollama Chat (`ollama_chat`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Follama) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Oobabooga (`oobabooga`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fopenai_compatible) | ✅ | ✅ | ✅ |  |  | ✅ | ✅ | ✅ | ✅ |  |\n| [OpenAI (`openai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fopenai) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |  |\n| [OpenAI-like (`openai_like`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fopenai_compatible) |  |  |  | ✅ |  |  |  |  |  |  |\n| [OpenRouter (`openrouter`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fopenrouter) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [OVHCloud AI Endpoints (`ovhcloud`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fovhcloud) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Perplexity AI (`perplexity`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fperplexity) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Petals (`petals`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fpetals) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Predibase (`predibase`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fpredibase) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Recraft (`recraft`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Frecraft) |  |  |  |  | ✅ |  |  |  |  |  |\n| [Replicate (`replicate`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Freplicate) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Sagemaker Chat (`sagemaker_chat`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Faws_sagemaker) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Sambanova (`sambanova`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fsambanova) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Snowflake (`snowflake`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fsnowflake) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Text Completion Codestral (`text-completion-codestral`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fcodestral) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Text Completion OpenAI (`text-completion-openai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ftext_completion_openai) | ✅ | ✅ | ✅ |  |  | ✅ | ✅ | ✅ | ✅ |  |\n| [Together AI (`together_ai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ftogetherai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Topaz (`topaz`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ftopaz) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Triton (`triton`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Ftriton-inference-server) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [V0 (`v0`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fv0) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Vercel AI Gateway (`vercel_ai_gateway`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fvercel_ai_gateway) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [VLLM (`vllm`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fvllm) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Volcengine (`volcengine`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fvolcano) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Voyage AI (`voyage`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fvoyage) |  |  |  | ✅ |  |  |  |  |  |  |\n| [WandB Inference (`wandb`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fwandb_inference) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Watsonx Text (`watsonx_text`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fwatsonx) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [xAI (`xai`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fxai) | ✅ | ✅ | ✅ |  |  |  |  |  |  |  |\n| [Xinference (`xinference`)](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders\u002Fxinference) |  |  |  | ✅ |  |  |  |  |  |  |\n\n[**阅读文档**](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002F)\n\n## 以开发者模式运行\n### 服务\n1. 在根目录设置 .env 文件\n2. 运行依赖服务 `docker-compose up db prometheus`\n\n### 后端\n1. （在根目录）创建虚拟环境 `python -m venv .venv`\n2. 激活虚拟环境 `source .venv\u002Fbin\u002Factivate`\n3. 安装依赖 `pip install -e \".[all]\"`\n4. `pip install prisma`\n5. `prisma generate`\n6. 启动代理后端 `python litellm\u002Fproxy\u002Fproxy_cli.py`\n\n### 前端\n1. 进入 `ui\u002Flitellm-dashboard`\n2. 安装依赖 `npm install`\n3. 运行 `npm run dev` 以启动仪表盘\n\n# 企业版\n面向需要更好安全性、用户管理和专业支持的公司\n\n[与创始人联系](https:\u002F\u002Fenterprise.litellm.ai\u002Fdemo)\n\n这包括：\n- ✅ **[LiteLLM 商业许可](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproxy\u002Fenterprise) 下的功能：**\n- ✅ **功能优先级排序**\n- ✅ **自定义集成**\n- ✅ **专业支持 - 专属 Discord + Slack**\n- ✅ **自定义 SLA（服务等级协议）**\n- ✅ **使用单点登录 (SSO) 的安全访问**\n\n# 贡献\n\n我们欢迎为 LiteLLM 做出贡献！无论您是修复错误、添加功能还是改进文档，我们都感谢您的帮助。\n\n## 贡献者快速开始\n\n这需要安装 Poetry。\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm.git\ncd litellm\nmake install-dev    # Install development dependencies\nmake format         # Format your code\nmake lint           # Run all linting checks\nmake test-unit      # Run unit tests\nmake format-check   # Check formatting only\n```\n\n有关详细的贡献指南，请参阅 [CONTRIBUTING.md](CONTRIBUTING.md)。\n\n## 代码质量 \u002F 代码检查\n\nLiteLLM 遵循 [Google Python 风格指南](https:\u002F\u002Fgoogle.github.io\u002Fstyleguide\u002Fpyguide.html)。\n\n我们的自动化检查包括：\n- **Black** 用于代码格式化\n- **Ruff** 用于代码检查和代码质量\n- **MyPy** 用于类型检查\n- **循环导入检测**\n- **导入安全检查**\n\n\n所有这些检查必须通过，您的 PR（拉取请求）才能被合并。\n\n\n# 支持 \u002F 与创始人交流\n\n- [预约演示 👋](https:\u002F\u002Fcalendly.com\u002Fd\u002F4mp-gd3-k5k\u002Fberriai-1-1-onboarding-litellm-hosted-version)\n- [社区 Discord 💭](https:\u002F\u002Fdiscord.gg\u002FwuPM9dRgDw)\n- [社区 Slack 💭](https:\u002F\u002Fwww.litellm.ai\u002Fsupport)\n- 我们的邮箱 ✉️ ishaan@berri.ai \u002F krrish@berri.ai\n\n# 我们为何构建此项目\n\n- **对简洁性的需求**：我们的代码在管理 Azure、OpenAI 和 Cohere 之间的调用及转换时变得极其复杂。\n\n# 贡献者\n\n\u003C!-- ALL-CONTRIBUTORS-LIST:START - Do not remove or modify this section -->\n\u003C!-- prettier-ignore-start -->\n\u003C!-- markdownlint-disable -->\n\n\u003C!-- markdownlint-restore -->\n\u003C!-- prettier-ignore-end -->\n\n\u003C!-- ALL-CONTRIBUTORS-LIST:END -->\n\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fgraphs\u002Fcontributors\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_readme_a2f9195dab1c.png\" \u002F>\n\u003C\u002Fa>","# LiteLLM 快速上手指南\n\nLiteLLM 是一个统一的接口库和代理服务器，允许您以 OpenAI 格式调用 100+ 个大语言模型（如 Bedrock、Azure、OpenAI、Anthropic、Groq 等）。它旨在简化多模型集成，提供路由、负载均衡和成本追踪功能。\n\n## 1. 环境准备\n\n*   **操作系统**: Linux, macOS, Windows\n*   **Python 版本**: 3.8 及以上\n*   **包管理器**: pip\n\n> 💡 **国内网络优化建议**：由于依赖包较大，建议在安装前配置国内镜像源以提升下载速度。\n> ```bash\n> pip config set global.index-url https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n> ```\n\n## 2. 安装步骤\n\n根据您的需求选择以下两种模式之一进行安装：\n\n### 模式一：Python SDK（直接集成到代码中）\n适用于开发者在 Python 项目中直接调用不同模型的 API。\n\n```bash\npip install litellm\n```\n\n### 模式二：AI Gateway（代理服务器）\n适用于需要集中管理密钥、监控流量或作为统一网关的场景。\n\n```bash\npip install 'litellm[proxy]'\n```\n\n## 3. 基本使用\n\nLiteLLM 支持两种主要使用方式：**Python SDK** 和 **Proxy Server**。\n\n### 方式 A：使用 Python SDK\n\n直接在代码中导入 `completion` 函数，设置环境变量中的 API Key，即可调用不同厂商的模型。\n\n```python\nfrom litellm import completion\nimport os\n\n# 设置您的 API Keys\nos.environ[\"OPENAI_API_KEY\"] = \"your-openai-key\"\nos.environ[\"ANTHROPIC_API_KEY\"] = \"your-anthropic-key\"\n\n# 调用 OpenAI 模型\nresponse = completion(model=\"openai\u002Fgpt-4o\", messages=[{\"role\": \"user\", \"content\": \"Hello!\"}])\n\n# 调用 Anthropic 模型\nresponse = completion(model=\"anthropic\u002Fclaude-sonnet-4-20250514\", messages=[{\"role\": \"user\", \"content\": \"Hello!\"}])\n```\n\n### 方式 B：使用 AI Gateway (Proxy Server)\n\n启动 LiteLLM 代理服务后，您可以像使用标准 OpenAI 客户端一样连接它。\n\n**1. 启动服务**\n```bash\nlitellm --model gpt-4o\n```\n\n**2. 通过 OpenAI 客户端连接**\n```python\nimport openai\n\nclient = openai.OpenAI(api_key=\"anything\", base_url=\"http:\u002F\u002F0.0.0.0:4000\")\nresponse = client.chat.completions.create(\n    model=\"gpt-4o\",\n    messages=[{\"role\": \"user\", \"content\": \"Hello!\"}]\n)\n```\n\n---\n\n更多高级功能（如 Agent 支持、MCP Tools 集成、路由策略等），请查阅官方文档：[LiteLLM Docs](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs)","某电商团队正在开发智能客服系统，需要同时接入 OpenAI、Anthropic 和 Azure 等多个大模型以平衡成本与效果。\n\n### 没有 litellm 时\n- 每个模型都需要单独编写适配代码，不同厂商的 API 返回格式不一致导致维护成本极高。\n- 想要切换更便宜的模型时，必须修改多处业务逻辑代码，测试和上线风险巨大。\n- 无法统一统计各模型的调用费用，财务部门难以进行精确的成本对账。\n- 缺乏统一的限流和熔断机制，高峰期容易触发第三方服务的频率限制导致服务中断。\n\n### 使用 litellm 后\n- litellm 提供统一的 OpenAI 格式接口，一套代码即可无缝调用所有支持的模型。\n- 通过配置文件即可动态切换后端模型，无需改动核心业务层代码，迭代效率大幅提升。\n- 内置成本追踪功能，自动记录每次调用的 Token 消耗和金额，实现精细化预算管理。\n- 部署 litellm 代理服务器支持负载均衡，自动分配请求避免单点故障并提升稳定性。\n\nlitellm 通过标准化接口和网关能力，显著降低了多模型集成的复杂度与管理成本。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FBerriAI_litellm_5b93aa3d.png","BerriAI","Berri AI","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002FBerriAI_499dcc04.png","The fastest way to take your LLM app to production",null,"founders@berri.ai","https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002F","https:\u002F\u002Fgithub.com\u002FBerriAI",[84,88,92,96,100,104,108,111,114,117],{"name":85,"color":86,"percentage":87},"Python","#3572A5",82.4,{"name":89,"color":90,"percentage":91},"TypeScript","#3178c6",15.8,{"name":93,"color":94,"percentage":95},"HTML","#e34c26",1.3,{"name":97,"color":98,"percentage":99},"JavaScript","#f1e05a",0.4,{"name":101,"color":102,"percentage":103},"Shell","#89e051",0.1,{"name":105,"color":106,"percentage":107},"Makefile","#427819",0,{"name":109,"color":110,"percentage":107},"Dockerfile","#384d54",{"name":112,"color":113,"percentage":107},"PowerShell","#012456",{"name":115,"color":116,"percentage":107},"Go","#00ADD8",{"name":118,"color":119,"percentage":107},"Ruby","#701516",42211,7001,"2026-04-05T10:39:52","NOASSERTION","未说明",{"notes":126,"python":124,"dependencies":127},"LiteLLM 主要作为 API 网关（Proxy Server）和 Python SDK 使用，支持调用 100+ 种外部大模型服务，无需本地运行模型或 GPU 环境。推荐使用 Docker 部署代理服务，支持虚拟密钥管理、多租户成本追踪、路由重试及日志记录功能。",[67,128,129,130,131],"openai","httpx","mcp","a2a",[13,26],[134,135,136,137,128,138,139,140,141,142,143,144,67,145],"anthropic","langchain","llm","llmops","ai-gateway","azure-openai","bedrock","gateway","openai-proxy","vertex-ai","llm-gateway","mcp-gateway",99,"2026-03-27T02:49:30.150509","2026-04-06T04:24:15.782265",[150,155,159,164,169,174],{"id":151,"question_zh":152,"answer_zh":153,"source_url":154},2755,"LiteLLM 代理长时间运行后内存占用过高且无法释放怎么办？","内存问题高度依赖于 LiteLLM 的配置以及所使用的端点和功能。如果遇到内存泄漏，请遵循 [故障排除指南](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Ftroubleshoot) 中的说明。建议检查是否使用了过多工具（tools\u002Ffunctions），尝试使用 gunicorn 并配置 worker 数量和 max_request 设置，同时关闭不必要的 spend logs 等监控功能。","https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fissues\u002F12685",{"id":156,"question_zh":157,"answer_zh":158,"source_url":154},2756,"自定义回调中报错“缺少用户 API 密钥、团队 ID 和用户 ID”如何解决？","该错误通常出现在 `proxy_track_cost_callback.py` 中。请确保在自定义回调配置中正确传递了 `user_api_key`、`team_id` 和 `user_id` 参数。检查调用链中这些关键信息是否在请求过程中被正确携带，缺失任一参数都会触发此异常。",{"id":160,"question_zh":161,"answer_zh":162,"source_url":163},2759,"LiteLLM 是否已兼容 OpenAI v1.0.0 SDK？","是的，LiteLLM v1.0.0 已在生产环境上线，完全兼容 OpenAI v1.0.0。同时项目也确保了向后兼容性，使用旧版 OpenAI SDK 的用户不会受到影响。可通过 `pip install --upgrade litellm==1.0.0.dev1` 获取开发版更新。","https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fissues\u002F774",{"id":165,"question_zh":166,"answer_zh":167,"source_url":168},2757,"Docker 部署 LiteLLM 时登录管理面板提示“认证错误，用户未找到”如何处理？","常见原因包括环境变量配置错误或旧数据冲突。1. 检查 `.env` 文件中的 `UI_USERNAME` 和 `UI_PASSWORD` 变量值开头是否有空格。2. 尝试删除旧的 Docker 卷（volumes），因为旧版本的配置可能导致冲突。注意：删除卷会清除所有现有配置和数据。建议将环境变量直接写入 `docker-compose.yml` 以确保正确加载。","https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fissues\u002F9264",{"id":170,"question_zh":171,"answer_zh":172,"source_url":173},2758,"使用 Claude-3 模型时流式输出（Streaming）不工作怎么办？","这可能是 Anthropic API 的临时性问题。如果遇到流式传输失败，可以先忽略并重新测试，通常随后会恢复正常。如果持续存在问题，请检查 API Key 有效性及网络状态，确认是否为服务端暂时故障。","https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fissues\u002F2314",{"id":175,"question_zh":176,"answer_zh":177,"source_url":154},2760,"如何排查 LiteLLM 代理的内存泄漏问题？","首先查看监控图表确认内存增长趋势。检查日志中是否有重复的错误堆栈（如 `Traceback`）。确认是否开启了过多的工具（tools）或函数调用，这会增加响应时间和内存消耗。建议对比直接使用 Azure API 与通过 LiteLLM 代理的差异，若代理延迟显著增加，需优化配置或减少并发请求数。",[179,184,189,194,199,204,209,214,219,224,229,234,238,243,248,253,258,263,268,273],{"id":180,"version":181,"summary_zh":182,"released_at":183},102238,"v1.83.1-nightly","This release was from the LiteLLM team. We are testing out a new signing process and it is safe to use. \r\n\r\nThis was a test release from us and co sign verify will not work for this release as we are testing a new cosign workflow from us. ","2026-04-03T05:24:22",{"id":185,"version":186,"summary_zh":187,"released_at":188},102239,"v1.83.0-nightly","## Verifying the Docker Image\r\n```\r\ncosign verify --insecure-ignore-tlog=true --key https:\u002F\u002Fraw.githubusercontent.com\u002FBerriAI\u002Flitellm\u002Fmain\u002Fcosign.pub \\\r\n  ghcr.io\u002Fberriai\u002Flitellm:v1.83.0-nightly\r\n```\r\nOutput:\r\n```\r\nVerification for ghcr.io\u002Fberriai\u002Flitellm:v1.83.0-nightly --\r\nThe following checks were performed on each of these signatures:\r\n  - The cosign claims were validated\r\n  - The signatures were verified against the specified public key\r\n```\r\n\r\n## What's Changed\r\n* ci: skip scheduled workflows on forks by @DmitriyAlergant in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24460\r\n* fix(proxy): ignore return_to in SSO when control_plane_url is not con… by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24475\r\n* Correct documentation of completion_model by @benlangfeld in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24413\r\n* chore: migrate route_preview.tsx from Tremor to Ant Design by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24485\r\n* docs: add audit log export to S3 documentation by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24486\r\n* fix(security_scans.sh): pin trivvy version by @krrish-berri-2 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24525\r\n* docs: add security update blog post for March 2026 supply chain incident by @ryan-crabbe-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24537\r\n* docs: minor updates to security update blog post by @ishaan-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24540\r\n* ci: remove publish-migrations and reset_stable workflows by @ishaan-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24541\r\n* security: remove .claude\u002Fsettings.json and block re-adding via semgrep by @ishaan-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24584\r\n* docs: remove phone numbers from readme and docs by @ishaan-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24587\r\n* Litellm ryan march 23 by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24490\r\n* [Infra] Pins GitHub Action workflows to specific hashes by @yuneng-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24594\r\n* docs(opencode): add guidance for dropping reasoningSummary param by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24468\r\n* [Infra] Pin GHA dependencies and remove unused load test files by @yuneng-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24607\r\n* Update README.md by @ishaan-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24649\r\n* [Infra] Migrate PyPI Publishing from CircleCI to GitHub Actions OIDC by @yuneng-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24654\r\n* Add zizmor to ci\u002Fcd  by @krrish-berri-2 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24663\r\n* feat(prometheus): add metrics for managed batch lifecycle by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24691\r\n* feat(openai): round-trip Responses API reasoning_items in chat completions by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24690\r\n* feat(fine-tuning): fix Azure OpenAI fine-tuning job creation by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24687\r\n* feat(gemini): add gemini-3.1-flash-live-preview to model cost map by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24665\r\n* feat(gemini): normalize AI Studio file retrieve URL by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24662\r\n* fix(anthropic): strip undocumented keys from metadata before sending to API by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24661\r\n* Litellm security townhall blog by @krrish-berri-2 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24692\r\n* fix(proxy): sanitize user_id input and block dangerous env var keys by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24624\r\n* feat(gemini): Lyria 3 preview models in cost map and docs by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24610\r\n* fix(openrouter): strip routing prefix for wildcard proxy deployments by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24603\r\n* fix(pricing): remove above_200k_tokens price tiers for claude-opus-4-6 and claude-sonnet-4-6 by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24689\r\n* fix(team-routing): preserve sibling deployment candidates for team public models by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24688\r\n* [Fix] Pin Prisma Node.js dependency in CI workflows by @yuneng-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24696\r\n* [Infra] Improve CodeQL scanning coverage and schedule by @yuneng-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24697\r\n* fix(proxy): enforce budget limits across multi-pod deployments via Redis-backed spend counters by @michelligabriele in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24682\r\n* [Infra] Automated schema.prisma sync and drift detection by @yuneng-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24705\r\n* fix: add \u002Fuser\u002Fbulk_update to management routes by @ryan-crabbe-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24708\r\n* fix(sso): pass decoded JWT access token to role mapping during SSO login by @ryan-crabbe-berri in https:\u002F\u002Fgithu","2026-03-31T21:49:27",{"id":190,"version":191,"summary_zh":192,"released_at":193},102240,"v1.82.6.rc.2","**Full Changelog**: https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fcompare\u002Fv1.82.6.dev1...v1.82.6.rc.2","2026-03-24T03:39:29",{"id":195,"version":196,"summary_zh":197,"released_at":198},102241,"v1.82.6.dev2","## What's Changed\n* fix(proxy): post-call guardrail response not captured for logging by @michelligabriele in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23910\n* feat(prometheus): include spend_logs_metadata in custom labels by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24434\n* feat(proxy): add project_alias tracking in callbacks by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24432\n\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fcompare\u002Fv1.82.5.dev.1...v1.82.6.dev2","2026-03-24T04:25:34",{"id":200,"version":201,"summary_zh":202,"released_at":203},102242,"v1.82.3-stable.patch.2","## What's Changed\n* fix: don't close HTTP\u002FSDK clients on LLMClientCache eviction by @ishaan-jaff in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F22926\n* [Feature] UI - Keys: Add Organization Dropdown to Create\u002FEdit Key by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23595\n* [Test] UI - Add Unit Tests for 5 Untested Components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23607\n* [Feature] UI - Internal Users: Add\u002FRemove Team Membership by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23638\n* [Feature] UI - Default Team Settings: Modernize page and fix defaults application by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23614\n* [Fix] Tag list endpoint 500 from invalid Prisma group_by kwargs by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23606\n* [Feature] UI - Usage: Auto-paginate daily spend data by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23622\n* [Infra] Merge internal dev 3\u002F13 with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23658\n* feat: add sagemaker_nova provider for Amazon Nova models on SageMaker by @ryanh-ai in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F21542\n* [Fix] Team Admin 403 on \u002Fuser\u002Ffilter\u002Fui With Org Scope Flag by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23671\n* fix: redact secrets from proxy log output by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23668\n* fix: prisma migrate deploy failures on pre-existing instances by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23655\n* [Fix] Hanging CI Tests in custom_httpx test_http_handler by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23674\n* Add CodSpeed continuous performance benchmarks by @codspeed-hq[bot] in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23676\n* [Infra] Merge internal dev with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23654\n* Litellm fix redaction filter logging v2 by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23667\n* [Fix] Bump PyJWT to ^2.12.0 for Security by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23678\n* Root cause fix - migrate all logging update to use 1 function - for centralized kwarg updates by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23659\n* fix: make db migration failure exit opt-in via --enforce_prisma_migration_check by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23675\n* [Infra] Optimize CI Pipeline by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23721\n\n## New Contributors\n* @codspeed-hq[bot] made their first contribution in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23676\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fcompare\u002Fv1.82.0.patch5...v1.82.3-stable.patch.2","2026-03-24T06:10:51",{"id":205,"version":206,"summary_zh":207,"released_at":208},102243,"v1.82.3-stable.patch.1","## What's Changed\n* fix: don't close HTTP\u002FSDK clients on LLMClientCache eviction by @ishaan-jaff in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F22926\n* [Feature] UI - Keys: Add Organization Dropdown to Create\u002FEdit Key by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23595\n* [Test] UI - Add Unit Tests for 5 Untested Components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23607\n* [Feature] UI - Internal Users: Add\u002FRemove Team Membership by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23638\n* [Feature] UI - Default Team Settings: Modernize page and fix defaults application by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23614\n* [Fix] Tag list endpoint 500 from invalid Prisma group_by kwargs by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23606\n* [Feature] UI - Usage: Auto-paginate daily spend data by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23622\n* [Infra] Merge internal dev 3\u002F13 with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23658\n* feat: add sagemaker_nova provider for Amazon Nova models on SageMaker by @ryanh-ai in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F21542\n* [Fix] Team Admin 403 on \u002Fuser\u002Ffilter\u002Fui With Org Scope Flag by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23671\n* fix: redact secrets from proxy log output by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23668\n* fix: prisma migrate deploy failures on pre-existing instances by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23655\n* [Fix] Hanging CI Tests in custom_httpx test_http_handler by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23674\n* Add CodSpeed continuous performance benchmarks by @codspeed-hq[bot] in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23676\n* [Infra] Merge internal dev with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23654\n* Litellm fix redaction filter logging v2 by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23667\n* [Fix] Bump PyJWT to ^2.12.0 for Security by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23678\n* Root cause fix - migrate all logging update to use 1 function - for centralized kwarg updates by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23659\n* fix: make db migration failure exit opt-in via --enforce_prisma_migration_check by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23675\n* [Infra] Optimize CI Pipeline by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23721\n\n## New Contributors\n* @codspeed-hq[bot] made their first contribution in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23676\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fcompare\u002Fv1.82.0.patch5...v1.82.3-stable.patch.1","2026-03-24T03:33:50",{"id":210,"version":211,"summary_zh":212,"released_at":213},102244,"v1.82.6.dev1","**Full Changelog**: https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fcompare\u002Fv1.82.6.rc.1...v1.82.6.dev1","2026-03-23T17:45:37",{"id":215,"version":216,"summary_zh":217,"released_at":218},102245,"v1.82.3.dev.5","**Full Changelog**: https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fcompare\u002Fv1.82.5.dev.1...v1.82.3.dev.5","2026-03-24T00:17:30",{"id":220,"version":221,"summary_zh":222,"released_at":223},102246,"v1.82.6.rc.1","## What's Changed\n* fix(anthropic): preserve thinking.summary when routing to OpenAI Responses API by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F21441\n* docs: add thinking.summary field to \u002Fv1\u002Fmessages and reasoning docs by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F22823\n* fix(gemini): resolve image token undercounting in usage metadata by @gustipardo in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F22608\n* feat(anthropic): add opt-out flag for default reasoning summary by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F22904\n* fix(anthropic): align translate_thinking_for_model with default summary injection + docs by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F22909\n* fix: don't close HTTP\u002FSDK clients on LLMClientCache eviction by @ishaan-jaff in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F22926\n* [Feature] UI - Keys: Add Organization Dropdown to Create\u002FEdit Key by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23595\n* [Test] UI - Add Unit Tests for 5 Untested Components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23607\n* [Feature] UI - Internal Users: Add\u002FRemove Team Membership by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23638\n* [Feature] UI - Default Team Settings: Modernize page and fix defaults application by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23614\n* [Fix] Tag list endpoint 500 from invalid Prisma group_by kwargs by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23606\n* [Feature] UI - Usage: Auto-paginate daily spend data by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23622\n* [Infra] Merge internal dev 3\u002F13 with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23658\n* feat: add sagemaker_nova provider for Amazon Nova models on SageMaker by @ryanh-ai in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F21542\n* [Fix] Team Admin 403 on \u002Fuser\u002Ffilter\u002Fui With Org Scope Flag by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23671\n* fix: redact secrets from proxy log output by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23668\n* fix: prisma migrate deploy failures on pre-existing instances by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23655\n* [Fix] Hanging CI Tests in custom_httpx test_http_handler by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23674\n* Add CodSpeed continuous performance benchmarks by @codspeed-hq[bot] in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23676\n* [Infra] Merge internal dev with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23654\n* Litellm fix redaction filter logging v2 by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23667\n* [Fix] Bump PyJWT to ^2.12.0 for Security by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23678\n* Root cause fix - migrate all logging update to use 1 function - for centralized kwarg updates by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23659\n* fix: make db migration failure exit opt-in via --enforce_prisma_migration_check by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23675\n* [Infra] Optimize CI Pipeline by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23721\n* Fix: Vertex ai Batch Output File Download Fails with 500 by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23718\n* docs(blog): add WebRTC blog post link by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23547\n* Refactor: Filtering beta header after transformation by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23715\n* fix(streaming): preserve custom attributes on final stream chunk by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23530\n* [Feat] Add create character endpoints and other new videos Endpoints by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23737\n* Litellm oss staging 03 14 2026 by @RheagalFire in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23686\n* fix: align DefaultInternalUserParams Pydantic default with runtime fallback  by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23666\n* [Test] UI Dashboard - Add unit tests for 5 untested files by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23773\n* [Infra] Merging RC Branch with Main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23786\n* [Fix] Privilege Escalation on \u002Fkey\u002Fblock, \u002Fkey\u002Funblock, and \u002Fkey\u002Fupdate max_budget by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23781\n* chore(ui): migrate DefaultUserSettings buttons from Tremor to antd by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23787\n* fix: set oauth2_flow when building MCPServer in _execute_with_mcp_client by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23468\n* [Fix] UI - Logs: Empty Filter Results Show Stale Data by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23792\n* Litellm update blog posts rss by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23791\n* [Fix] Prevent Internal Users from Creating Invalid Keys by @yuneng-jian","2026-03-23T16:27:11",{"id":225,"version":226,"summary_zh":227,"released_at":228},102247,"v1.82.6.dev.1","## What's Changed\n* [Test] UI: Add unit tests for 10 untested components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23891\n* fix: auto-recover shared aiohttp session when closed by @voidborne-d in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23808\n* [Feature] \u002Fv2\u002Fteam\u002Flist: Add org admin access control, members_count, and indexes by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23938\n* [Refactor] UI - Playground: Extract FilePreviewCard from ChatUI by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23973\n* docs: add v1.82.3 release notes by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23820\n* fix(proxy): model-level guardrails not executing for non-streaming post_call by @michelligabriele in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23774\n* fix(proxy): prevent duplicate callback logs for pass-through endpoint failures by @michelligabriele in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23509\n* docs: Revamp documentation site with new navigation, landing pages, and styling by @Arindam200 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24023\n* Fix langfuse otel traceparent propagation by @jyeros in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24048\n* [Test] UI: Add unit tests for 10 untested components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24036\n* [Fix] UI - Logs: Guardrail Mode Type Crash on Non-String Values by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24035\n* [Staging] - Ishaan March 17th  by @ishaan-jaff in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23903\n* [Infra] Merge daily branch with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24055\n* [Fix] UI - Default Team Settings: Add Missing Permission Options by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24039\n* fix: \u002Fkey\u002Fblock and \u002Fkey\u002Funblock return 404 (not 401) for non-existent keys by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23977\n* [Refactor] UI - Playground: Extract ChatMessageBubble from ChatUI by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24062\n* [Fix] Key Update Endpoint Returns 401 Instead of 404 for Nonexistent Keys by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24063\n* fix: surface Anthropic code execution results as code_interpreter_call in Responses API by @andrzej-pomirski-yohana in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23784\n* [Feature] UI - Leftnav: Add external link icon to Learning Resources by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24069\n* Fix thinking blocks dropped when thinking field is null by @xr843 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24070\n* Preserve router model_group in generic API logs by @emerzon in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24044\n* Fix\u002Fproxy only failure call type by @alilxxey in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24050\n* fix(langsmith): populate usage_metadata in outputs for Cost column by @themavik in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24043\n* Fix model repetition detection performance by @hytromo in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F18120\n* fix: fix logging for response incomplete streaming + custom pricing on \u002Fv1\u002Fmessages and \u002Fv1\u002Fresponses  by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24080\n* docs: add Contributing to Guardrails section to Guardrail Providers sidebar by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24083\n* chore: regenerate poetry.lock to unblock CI (pyproject.toml content hash drift) by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24082\n* Ishaan - March 18th changes  by @ishaan-jaff in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24056\n* [Test] UI: Add vitest coverage for 10 untested components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24144\n* chore: apply black formatting to fix lint CI by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24092\n* feat(proxy): use AZURE_DEFAULT_API_VERSION for proxy --api_version default by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24120\n* chore: apply black formatting to fix lint CI (batch 2) by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24153\n* [Refactor] UI - API Reference: Migrate to Path-Based Routing by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24155\n* chore: add poetry check --lock to lint CI to prevent stale lockfile merges by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24159\n* [Infra] Merge daily internal branch with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24150\n* chore: apply black formatting to proxy\u002F_types.py to fix lint CI by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24167\n* docs: add LiteLLM license key environment variable instructions by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24168\n* Litellm oss staging 03 18 2026 by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24081\n* chore: apply black formatting to fix lint CI (batch 3) by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24173\n* docs(release-notes","2026-03-22T07:16:11",{"id":230,"version":231,"summary_zh":232,"released_at":233},102248,"v1.82.6-nightly","## What's Changed\n* fix(anthropic): preserve thinking.summary when routing to OpenAI Responses API by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F21441\n* docs: add thinking.summary field to \u002Fv1\u002Fmessages and reasoning docs by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F22823\n* fix(gemini): resolve image token undercounting in usage metadata by @gustipardo in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F22608\n* feat(anthropic): add opt-out flag for default reasoning summary by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F22904\n* fix(anthropic): align translate_thinking_for_model with default summary injection + docs by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F22909\n* fix(fireworks): skip #transform=inline for base64 data URLs by @awais786 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23729\n* fix(langsmith): avoid no running event loop during sync init by @pandego in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23727\n* fix(gemini): support images in tool_results for \u002Fv1\u002Fmessages routing by @awais786 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23724\n* fix(model-prices): correct supported_regions for Vertex AI DeepSeek models by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23864\n* fix(model-prices): restore gpt-4-0314 by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23753\n* fix(cache): Fix Redis cluster caching by @cohml in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23480\n* fix(proxy): convert max_budget to float when set via environment variable by @rstar327 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23855\n* Add Akto Guardrails to LiteLLM by @rzeta-10 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23250\n* fix(core): map Anthropic 'refusal' finish reason to 'content_filter' by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23899\n* fix(vertex): streaming finish_reason='stop' instead of 'tool_calls' for gemini-3.1-flash-lite-preview by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23895\n* fix: map Chat Completion file type to Responses API input_file by @gambletan in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23618\n* fix(vertex): respect vertex_count_tokens_location for Claude count_tokens by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23907\n* fix(anthropic): preserve cache directive on file-type content blocks by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23906\n* fix(mistral): preserve diarization segments in transcription response by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23925\n* fix(gemini): pass model to context caching URL builder for custom api_base by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23928\n* fix(azure): auto-route gpt-5.4+ tools+reasoning to Responses API by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23926\n* merge main by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24119\n* [Test] UI: Add vitest coverage for 10 untested components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24144\n* [Refactor] UI - API Reference: Migrate to Path-Based Routing by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24155\n* [Fix] UI: AntD Messages Not Rendering by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24192\n* [Feature] UI - Teams: Modernize Teams Table by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24189\n* [Refactor] Extract useChatHistory hook from ChatUI.tsx by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24172\n* Day 0: gpt 5.4 mini and nano support by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23958\n* [feat]Add prompt management support for responses api by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23999\n* docs(vertex): add PayGo\u002FPriority tutorial and cost tracking flow diagramLitellm vertex paygo tutorial by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24009\n* feat(file_search): align emulated Responses behavior with native output by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23969\n* fix(proxy): pass-through subpath auth for non-admin users by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24079\n* fix(polling): check rate limits before creating polling ID by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24106\n* docs: note min version for encrypted_content_affinity by @milan-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23854\n* feat(router): add per-model-group deployment affinity by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24110\n* Litellm fix ensure alternating roles by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24015\n* fix(bedrock): respect api_base and aws_bedrock_runtime_endpoint in count_tokens endpoint by @stias in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24199\n* fix(gemini-embeddings): convert task_type to camelCase taskType for Gemini API by @christopherbaer in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24191\n* fix(vertex-ai): support batch cancel via Vertex API  by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23957\n* fix(azure-ai-agents): preserve annotat","2026-03-22T07:00:36",{"id":235,"version":236,"summary_zh":232,"released_at":237},102249,"v1.82.5.dev.1","2026-03-23T18:49:50",{"id":239,"version":240,"summary_zh":241,"released_at":242},102250,"v1.82.1.dev.1","**Full Changelog**: https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fcompare\u002Fv1.82.1-nightly...v1.82.1.dev.1","2026-03-22T03:11:14",{"id":244,"version":245,"summary_zh":246,"released_at":247},102251,"v1.81.14.dev.3","## What's Changed\n* [Test] UI: Add vitest coverage for 10 untested components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24144\n* [Refactor] UI - API Reference: Migrate to Path-Based Routing by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24155\n* [Fix] UI: AntD Messages Not Rendering by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24192\n* [Feature] UI - Teams: Modernize Teams Table by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24189\n* [Refactor] Extract useChatHistory hook from ChatUI.tsx by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24172\n* fix(bedrock): respect api_base and aws_bedrock_runtime_endpoint in count_tokens endpoint by @stias in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24199\n* fix(gemini-embeddings): convert task_type to camelCase taskType for Gemini API by @christopherbaer in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24191\n* fix(logging): merge hidden_params into metadata for streaming requests by @milan-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24220\n* docs(pricing): add official source links for Azure DeepSeek & Cohere … by @Jayachander123 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F20181\n* fix(transcription): fix cost_per_second calculation for audio transcription models by @gvioss in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23842\n* fix(moonshot): preserve reasoning_content on Pydantic Message objects in multi-turn tool calls by @BillionClaw in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23828\n* fix(proxy): add team_member_budget_duration to NewTeamRequest by @michelligabriele in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23484\n* [Feature] Team MCP Server Manager Role by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24171\n* [Docs] Add missing team_member_budget_duration param to new_team docstring by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24243\n* feat: add control plane for multi-proxy worker management by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24217\n* Revert \"[Feature] Team MCP Server Manager Role\" by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24255\n* fix(test): mock get_auth_header instead of get_api_key in anthropic file content test by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24258\n* [Infra] Merge internal dev branch with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24260\n* fix(adapter): add additionalProperties: false for OpenAI strict mode in Anthropic adapter by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24072\n* feat(gemini): support context circulation for server-side tool combination by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24073\n* feat(bedrock): support cache_control_injection_points for tool_config location by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24076\n* Revert \"fix(transcription): fix cost_per_second calculation for audio transcription models\" by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24297\n* Litellm oss staging 03 19 2026 by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24174\n* docs: add Gemini\u002FVertex AI prompt caching support to docs by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24222\n* Litellm ishaan march 20 by @ishaan-jaff in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24303\n\n## New Contributors\n* @stias made their first contribution in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24199\n* @christopherbaer made their first contribution in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24191\n* @Jayachander123 made their first contribution in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F20181\n* @BillionClaw made their first contribution in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23828\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fcompare\u002Fv1.82.3.dev.2...v1.81.14.dev.3","2026-03-22T01:13:29",{"id":249,"version":250,"summary_zh":251,"released_at":252},102252,"v1.82.3.dev.4","## What's Changed\n* fix: don't close HTTP\u002FSDK clients on LLMClientCache eviction by @ishaan-jaff in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F22926\n* [Feature] UI - Keys: Add Organization Dropdown to Create\u002FEdit Key by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23595\n* [Test] UI - Add Unit Tests for 5 Untested Components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23607\n* [Feature] UI - Internal Users: Add\u002FRemove Team Membership by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23638\n* [Feature] UI - Default Team Settings: Modernize page and fix defaults application by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23614\n* [Fix] Tag list endpoint 500 from invalid Prisma group_by kwargs by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23606\n* [Feature] UI - Usage: Auto-paginate daily spend data by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23622\n* [Infra] Merge internal dev 3\u002F13 with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23658\n* feat: add sagemaker_nova provider for Amazon Nova models on SageMaker by @ryanh-ai in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F21542\n* [Fix] Team Admin 403 on \u002Fuser\u002Ffilter\u002Fui With Org Scope Flag by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23671\n* fix: redact secrets from proxy log output by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23668\n* fix: prisma migrate deploy failures on pre-existing instances by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23655\n* [Fix] Hanging CI Tests in custom_httpx test_http_handler by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23674\n* Add CodSpeed continuous performance benchmarks by @codspeed-hq[bot] in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23676\n* [Infra] Merge internal dev with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23654\n* Litellm fix redaction filter logging v2 by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23667\n* [Fix] Bump PyJWT to ^2.12.0 for Security by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23678\n* Root cause fix - migrate all logging update to use 1 function - for centralized kwarg updates by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23659\n* fix: make db migration failure exit opt-in via --enforce_prisma_migration_check by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23675\n* [Infra] Optimize CI Pipeline by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23721\n\n## New Contributors\n* @codspeed-hq[bot] made their first contribution in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23676\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fcompare\u002Fv1.82.0.patch5...v1.82.3.dev.4","2026-03-21T06:19:51",{"id":254,"version":255,"summary_zh":256,"released_at":257},102253,"v1.81.14.dev.2","## What's Changed\n* [Test] UI: Add vitest coverage for 10 untested components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24144\n* [Refactor] UI - API Reference: Migrate to Path-Based Routing by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24155\n* [Fix] UI: AntD Messages Not Rendering by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24192\n* [Feature] UI - Teams: Modernize Teams Table by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24189\n* [Refactor] Extract useChatHistory hook from ChatUI.tsx by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24172\n* fix(bedrock): respect api_base and aws_bedrock_runtime_endpoint in count_tokens endpoint by @stias in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24199\n* fix(gemini-embeddings): convert task_type to camelCase taskType for Gemini API by @christopherbaer in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24191\n* fix(logging): merge hidden_params into metadata for streaming requests by @milan-berri in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24220\n* docs(pricing): add official source links for Azure DeepSeek & Cohere … by @Jayachander123 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F20181\n* fix(transcription): fix cost_per_second calculation for audio transcription models by @gvioss in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23842\n* fix(moonshot): preserve reasoning_content on Pydantic Message objects in multi-turn tool calls by @BillionClaw in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23828\n* fix(proxy): add team_member_budget_duration to NewTeamRequest by @michelligabriele in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23484\n* [Feature] Team MCP Server Manager Role by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24171\n* [Docs] Add missing team_member_budget_duration param to new_team docstring by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24243\n* feat: add control plane for multi-proxy worker management by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24217\n* Revert \"[Feature] Team MCP Server Manager Role\" by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24255\n* fix(test): mock get_auth_header instead of get_api_key in anthropic file content test by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24258\n* [Infra] Merge internal dev branch with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24260\n* fix(adapter): add additionalProperties: false for OpenAI strict mode in Anthropic adapter by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24072\n* feat(gemini): support context circulation for server-side tool combination by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24073\n* feat(bedrock): support cache_control_injection_points for tool_config location by @Chesars in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24076\n* Revert \"fix(transcription): fix cost_per_second calculation for audio transcription models\" by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24297\n* Litellm oss staging 03 19 2026 by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24174\n* docs: add Gemini\u002FVertex AI prompt caching support to docs by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24222\n* Litellm ishaan march 20 by @ishaan-jaff in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24303\n\n## New Contributors\n* @stias made their first contribution in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24199\n* @christopherbaer made their first contribution in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24191\n* @Jayachander123 made their first contribution in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F20181\n* @BillionClaw made their first contribution in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23828\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fcompare\u002Fv1.82.3.dev.2...v1.81.14.dev.2","2026-03-21T23:38:31",{"id":259,"version":260,"summary_zh":261,"released_at":262},102254,"v1.82.5-nightly","## What's Changed\n* [Test] UI: Add unit tests for 10 untested components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23891\n* fix: auto-recover shared aiohttp session when closed by @voidborne-d in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23808\n* [Feature] \u002Fv2\u002Fteam\u002Flist: Add org admin access control, members_count, and indexes by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23938\n* [Refactor] UI - Playground: Extract FilePreviewCard from ChatUI by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23973\n* docs: add v1.82.3 release notes by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23820\n* fix(proxy): model-level guardrails not executing for non-streaming post_call by @michelligabriele in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23774\n* fix(proxy): prevent duplicate callback logs for pass-through endpoint failures by @michelligabriele in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23509\n* docs: Revamp documentation site with new navigation, landing pages, and styling by @Arindam200 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24023\n* Fix langfuse otel traceparent propagation by @jyeros in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24048\n* [Test] UI: Add unit tests for 10 untested components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24036\n* [Fix] UI - Logs: Guardrail Mode Type Crash on Non-String Values by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24035\n* [Staging] - Ishaan March 17th  by @ishaan-jaff in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23903\n* [Infra] Merge daily branch with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24055\n* [Fix] UI - Default Team Settings: Add Missing Permission Options by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24039\n* fix: \u002Fkey\u002Fblock and \u002Fkey\u002Funblock return 404 (not 401) for non-existent keys by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23977\n* [Refactor] UI - Playground: Extract ChatMessageBubble from ChatUI by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24062\n* [Fix] Key Update Endpoint Returns 401 Instead of 404 for Nonexistent Keys by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24063\n* fix: surface Anthropic code execution results as code_interpreter_call in Responses API by @andrzej-pomirski-yohana in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23784\n* [Feature] UI - Leftnav: Add external link icon to Learning Resources by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24069\n* Fix thinking blocks dropped when thinking field is null by @xr843 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24070\n* Preserve router model_group in generic API logs by @emerzon in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24044\n* Fix\u002Fproxy only failure call type by @alilxxey in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24050\n* fix(langsmith): populate usage_metadata in outputs for Cost column by @themavik in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24043\n* Fix model repetition detection performance by @hytromo in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F18120\n* fix: fix logging for response incomplete streaming + custom pricing on \u002Fv1\u002Fmessages and \u002Fv1\u002Fresponses  by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24080\n* docs: add Contributing to Guardrails section to Guardrail Providers sidebar by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24083\n* chore: regenerate poetry.lock to unblock CI (pyproject.toml content hash drift) by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24082\n* Ishaan - March 18th changes  by @ishaan-jaff in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24056\n* chore: apply black formatting to fix lint CI by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24092\n* feat(proxy): use AZURE_DEFAULT_API_VERSION for proxy --api_version default by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24120\n* chore: apply black formatting to fix lint CI (batch 2) by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24153\n* chore: add poetry check --lock to lint CI to prevent stale lockfile merges by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24159\n* [Infra] Merge daily internal branch with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24150\n* chore: apply black formatting to proxy\u002F_types.py to fix lint CI by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24167\n* docs: add LiteLLM license key environment variable instructions by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24168\n* Litellm oss staging 03 18 2026 by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24081\n* chore: apply black formatting to fix lint CI (batch 3) by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24173\n* docs(release-notes): complete v1.82.3 changelog with full PR audit by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24090\n* feat(anthropic): support ANTHROPIC_AUTH_TOKEN and ANTHROPIC_BASE_URL env vars by @devin-petersohn in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F2","2026-03-21T00:28:00",{"id":264,"version":265,"summary_zh":266,"released_at":267},102255,"v1.82.3.dev.3","## What's Changed\n* [Feature] UI - Keys: Add Organization Dropdown to Create\u002FEdit Key by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23595\n* [Test] UI - Add Unit Tests for 5 Untested Components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23607\n* [Feature] UI - Internal Users: Add\u002FRemove Team Membership by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23638\n* [Feature] UI - Default Team Settings: Modernize page and fix defaults application by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23614\n* [Fix] Tag list endpoint 500 from invalid Prisma group_by kwargs by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23606\n* [Feature] UI - Usage: Auto-paginate daily spend data by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23622\n* [Infra] Merge internal dev 3\u002F13 with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23658\n* feat: add sagemaker_nova provider for Amazon Nova models on SageMaker by @ryanh-ai in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F21542\n* [Fix] Team Admin 403 on \u002Fuser\u002Ffilter\u002Fui With Org Scope Flag by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23671\n* fix: redact secrets from proxy log output by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23668\n* fix: prisma migrate deploy failures on pre-existing instances by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23655\n* [Fix] Hanging CI Tests in custom_httpx test_http_handler by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23674\n* Add CodSpeed continuous performance benchmarks by @codspeed-hq[bot] in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23676\n* [Infra] Merge internal dev with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23654\n* Litellm fix redaction filter logging v2 by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23667\n* [Fix] Bump PyJWT to ^2.12.0 for Security by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23678\n* Root cause fix - migrate all logging update to use 1 function - for centralized kwarg updates by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23659\n* fix: make db migration failure exit opt-in via --enforce_prisma_migration_check by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23675\n* [Infra] Optimize CI Pipeline by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23721\n* Fix: Vertex ai Batch Output File Download Fails with 500 by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23718\n* docs(blog): add WebRTC blog post link by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23547\n* Refactor: Filtering beta header after transformation by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23715\n* fix(streaming): preserve custom attributes on final stream chunk by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23530\n* [Feat] Add create character endpoints and other new videos Endpoints by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23737\n* Litellm oss staging 03 14 2026 by @RheagalFire in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23686\n* fix: align DefaultInternalUserParams Pydantic default with runtime fallback  by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23666\n* [Test] UI Dashboard - Add unit tests for 5 untested files by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23773\n* [Infra] Merging RC Branch with Main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23786\n* [Fix] Privilege Escalation on \u002Fkey\u002Fblock, \u002Fkey\u002Funblock, and \u002Fkey\u002Fupdate max_budget by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23781\n* chore(ui): migrate DefaultUserSettings buttons from Tremor to antd by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23787\n* fix: set oauth2_flow when building MCPServer in _execute_with_mcp_client by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23468\n* [Fix] UI - Logs: Empty Filter Results Show Stale Data by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23792\n* Litellm update blog posts rss by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23791\n* [Fix] Prevent Internal Users from Creating Invalid Keys by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23795\n* [Fix] Key Alias Re-validation on Update Blocks Legacy Aliases by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23798\n* fix: Register DynamoAI guardrail initializer and enum entry by @Harshit28j in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23752\n* docs: add v1.82.3 release notes by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23816\n* Revert \"docs: add v1.82.3 release notes\" by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23817\n* [Feature] Disable Custom Virtual Key Values via UI Setting by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23812\n* fix(ui): CSV export empty on Global Usage page by @ryan-crabbe in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23819\n* fix: langfuse trace leak key on model params by @Harsh","2026-03-21T00:17:18",{"id":269,"version":270,"summary_zh":271,"released_at":272},102256,"v1.82.3.dev.2","## What's Changed\n* [Test] UI: Add unit tests for 10 untested components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23891\n* fix: auto-recover shared aiohttp session when closed by @voidborne-d in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23808\n* [Feature] \u002Fv2\u002Fteam\u002Flist: Add org admin access control, members_count, and indexes by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23938\n* [Refactor] UI - Playground: Extract FilePreviewCard from ChatUI by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23973\n* docs: add v1.82.3 release notes by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23820\n* fix(proxy): model-level guardrails not executing for non-streaming post_call by @michelligabriele in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23774\n* fix(proxy): prevent duplicate callback logs for pass-through endpoint failures by @michelligabriele in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23509\n* docs: Revamp documentation site with new navigation, landing pages, and styling by @Arindam200 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24023\n* Fix langfuse otel traceparent propagation by @jyeros in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24048\n* [Test] UI: Add unit tests for 10 untested components by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24036\n* [Fix] UI - Logs: Guardrail Mode Type Crash on Non-String Values by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24035\n* [Staging] - Ishaan March 17th  by @ishaan-jaff in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23903\n* [Infra] Merge daily branch with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24055\n* [Fix] UI - Default Team Settings: Add Missing Permission Options by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24039\n* fix: \u002Fkey\u002Fblock and \u002Fkey\u002Funblock return 404 (not 401) for non-existent keys by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23977\n* [Refactor] UI - Playground: Extract ChatMessageBubble from ChatUI by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24062\n* [Fix] Key Update Endpoint Returns 401 Instead of 404 for Nonexistent Keys by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24063\n* fix: surface Anthropic code execution results as code_interpreter_call in Responses API by @andrzej-pomirski-yohana in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23784\n* [Feature] UI - Leftnav: Add external link icon to Learning Resources by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24069\n* Fix thinking blocks dropped when thinking field is null by @xr843 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24070\n* Preserve router model_group in generic API logs by @emerzon in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24044\n* Fix\u002Fproxy only failure call type by @alilxxey in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24050\n* fix(langsmith): populate usage_metadata in outputs for Cost column by @themavik in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24043\n* Fix model repetition detection performance by @hytromo in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F18120\n* fix: fix logging for response incomplete streaming + custom pricing on \u002Fv1\u002Fmessages and \u002Fv1\u002Fresponses  by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24080\n* docs: add Contributing to Guardrails section to Guardrail Providers sidebar by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24083\n* chore: regenerate poetry.lock to unblock CI (pyproject.toml content hash drift) by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24082\n* fix(openai): handle missing 'id' field in streaming chunks for MiniMax by @chengyongru in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F23931\n* Ishaan - March 18th changes  by @ishaan-jaff in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24056\n* chore: apply black formatting to fix lint CI by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24092\n* feat(proxy): use AZURE_DEFAULT_API_VERSION for proxy --api_version default by @Sameerlite in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24120\n* chore: apply black formatting to fix lint CI (batch 2) by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24153\n* chore: add poetry check --lock to lint CI to prevent stale lockfile merges by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24159\n* [Infra] Merge daily internal branch with main by @yuneng-jiang in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24150\n* chore: apply black formatting to proxy\u002F_types.py to fix lint CI by @joereyna in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24167\n* docs: add LiteLLM license key environment variable instructions by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24168\n* fix(proxy): defer logging until post-call guardrails complete by @michelligabriele in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24135\n* Litellm oss staging 03 18 2026 by @krrishdholakia in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24081\n* Add IF NOT EXISTS to index creation in migration by @superpoussin22 in https:\u002F\u002Fgithub.com\u002FBerriAI\u002Flitellm\u002Fpull\u002F24105\n* f","2026-03-20T11:17:44",{"id":274,"version":275,"summary_zh":261,"released_at":276},102257,"v1.81.14.dev.1","2026-03-20T23:25:26"]