[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tool-AgentOps-AI--agentops":3,"similar-AgentOps-AI--agentops":229},{"id":4,"github_repo":5,"name":6,"description_en":7,"description_zh":8,"ai_summary_zh":9,"readme_en":10,"readme_zh":11,"quickstart_zh":12,"use_case_zh":13,"hero_image_url":14,"owner_login":15,"owner_name":16,"owner_avatar_url":17,"owner_bio":18,"owner_company":19,"owner_location":19,"owner_email":19,"owner_twitter":20,"owner_website":21,"owner_url":22,"languages":23,"stars":61,"forks":62,"last_commit_at":63,"license":64,"difficulty_score":65,"env_os":66,"env_gpu":66,"env_ram":66,"env_deps":67,"category_tags":75,"github_topics":81,"view_count":65,"oss_zip_url":19,"oss_zip_packed_at":19,"status":95,"created_at":96,"updated_at":97,"faqs":98,"releases":128},2833,"AgentOps-AI\u002Fagentops","agentops","Python SDK for AI agent monitoring, LLM cost tracking, benchmarking, and more. Integrates with most LLMs and agent frameworks including CrewAI, Agno, OpenAI Agents SDK, Langchain, Autogen, AG2, and CamelAI","AgentOps 是一款专为 AI 智能体（AI Agents）打造的开源可观测性与开发工具平台。它通过简单的 Python SDK，帮助开发者轻松监控智能体的运行状态、追踪大模型（LLM）的调用成本，并进行性能基准测试。\n\n在开发复杂的 AI 应用时，智能体的决策过程往往像“黑盒”，难以调试且成本不可控。AgentOps 正是为了解决这一痛点而生，它将智能体从原型设计到生产部署的全生命周期数据可视化，让开发者能清晰看到每一步的执行细节、延迟情况和费用支出，从而快速定位问题并优化效果。\n\n这款工具非常适合正在构建或维护 AI 智能体的开发者、研究人员及技术团队使用。无论您是使用 CrewAI、LangChain、AutoGen (AG2)、OpenAI Agents SDK 还是 CamelAI 等主流框架，AgentOps 都能无缝集成，无需大幅修改现有代码即可启用监控功能。\n\n作为 MIT 许可的开源项目，AgentOps 不仅提供了强大的云端仪表盘，其核心代码也完全开放，支持社区共同迭代。如果您希望提升 AI 智能体的稳定性、透明度并有效控制成本，AgentOps 是一个值得尝试","AgentOps 是一款专为 AI 智能体（AI Agents）打造的开源可观测性与开发工具平台。它通过简单的 Python SDK，帮助开发者轻松监控智能体的运行状态、追踪大模型（LLM）的调用成本，并进行性能基准测试。\n\n在开发复杂的 AI 应用时，智能体的决策过程往往像“黑盒”，难以调试且成本不可控。AgentOps 正是为了解决这一痛点而生，它将智能体从原型设计到生产部署的全生命周期数据可视化，让开发者能清晰看到每一步的执行细节、延迟情况和费用支出，从而快速定位问题并优化效果。\n\n这款工具非常适合正在构建或维护 AI 智能体的开发者、研究人员及技术团队使用。无论您是使用 CrewAI、LangChain、AutoGen (AG2)、OpenAI Agents SDK 还是 CamelAI 等主流框架，AgentOps 都能无缝集成，无需大幅修改现有代码即可启用监控功能。\n\n作为 MIT 许可的开源项目，AgentOps 不仅提供了强大的云端仪表盘，其核心代码也完全开放，支持社区共同迭代。如果您希望提升 AI 智能体的稳定性、透明度并有效控制成本，AgentOps 是一个值得尝试的专业助手。","\u003Cdiv align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fagentops.ai?ref=gh\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_1771207e29e0.png\" alt=\"Logo\">\n  \u003C\u002Fa>\n\u003C\u002Fdiv>\n\n\u003Cdiv align=\"center\">\n  \u003Cem>Observability and DevTool platform for AI Agents\u003C\u002Fem>\n\u003C\u002Fdiv>\n\n\u003Cbr \u002F>\n\n\u003Cdiv align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fpepy.tech\u002Fproject\u002Fagentops\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_ae8fe72d2972.png\" alt=\"Downloads\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fagentops-ai\u002Fagentops\u002Fissues\">\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fcommit-activity\u002Fm\u002Fagentops-ai\u002Fagentops\" alt=\"git commit activity\">\n  \u003C\u002Fa>\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fagentops?&color=3670A0\" alt=\"PyPI - Version\">\n  \u003Ca href=\"https:\u002F\u002Fopensource.org\u002Flicenses\u002FMIT\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-yellow.svg?&color=3670A0\" alt=\"License: MIT\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fsmithery.ai\u002Fserver\u002F@AgentOps-AI\u002Fagentops-mcp\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_2ab22774b7ca.png\"\u002F>\n  \u003C\u002Fa>\n\u003C\u002Fdiv>\n\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Ftwitter.com\u002Fagentopsai\u002F\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Ftwitter\u002Ffollow\u002Fagentopsai?style=social\" alt=\"Twitter\" style=\"height: 20px;\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FFagdcwwXRR\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fdiscord-7289da.svg?style=flat-square&logo=discord\" alt=\"Discord\" style=\"height: 20px;\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai\u002F?ref=gh\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDashboard-blue.svg?style=flat-square\" alt=\"Dashboard\" style=\"height: 20px;\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fintroduction\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDocumentation-orange.svg?style=flat-square\" alt=\"Documentation\" style=\"height: 20px;\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fentelligence.ai\u002FAgentOps-AI&agentops\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FChat%20with%20Docs-green.svg?style=flat-square\" alt=\"Chat with Docs\" style=\"height: 20px;\">\n  \u003C\u002Fa>\n\u003C\u002Fp>\n\n\u003Cdiv align=\"center\">\n  \u003Cvideo src=\"https:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002Fdfb4fa8d-d8c4-4965-9ff6-5b8514c1c22f\" width=\"650\" autoplay loop muted>\u003C\u002Fvideo>\n\u003C\u002Fdiv>\n\n\u003Cbr\u002F>\n\nAgentOps helps developers build, evaluate, and monitor AI agents. From prototype to production.\n\n## Open Source\n\nThe AgentOps app is open source under the MIT license. Explore the code in our [app directory](https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Ftree\u002Fmain\u002Fapp).\n\n## Key Integrations 🔌\n\n\u003Cdiv align=\"center\" style=\"background-color: white; padding: 20px; border-radius: 10px; margin: 0 auto; max-width: 800px;\">\n  \u003Cdiv style=\"display: flex; flex-wrap: wrap; justify-content: center; align-items: center; gap: 30px; margin-bottom: 20px;\">\n    \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fv2\u002Fintegrations\u002Fopenai_agents_python\">\u003Cimg src=\"docs\u002Fimages\u002Fexternal\u002Fopenai\u002Fagents-sdk.svg\" height=\"45\" alt=\"OpenAI Agents SDK\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcrewai\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_359269f184a0.png\" height=\"45\" alt=\"CrewAI\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.ag2.ai\u002Fdocs\u002Fecosystem\u002Fagentops\">\u003Cimg src=\"docs\u002Fimages\u002Fexternal\u002Fag2\u002Fag2-logo.svg\" height=\"45\" alt=\"AG2 (AutoGen)\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fmicrosoft\">\u003Cimg src=\"docs\u002Fimages\u002Fexternal\u002Fmicrosoft\u002Fmicrosoft_logo.svg\" height=\"45\" alt=\"Microsoft\">\u003C\u002Fa>\n  \u003C\u002Fdiv>\n  \n  \u003Cdiv style=\"display: flex; flex-wrap: wrap; justify-content: center; align-items: center; gap: 30px; margin-bottom: 20px;\">\n    \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Flangchain\">\u003Cimg src=\"docs\u002Fimages\u002Fexternal\u002Flangchain\u002Flangchain-logo.svg\" height=\"45\" alt=\"LangChain\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcamel\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_416126dc4ed3.png\" height=\"45\" alt=\"Camel AI\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.llamaindex.ai\u002Fen\u002Fstable\u002Fmodule_guides\u002Fobservability\u002F?h=agentops#agentops\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_2485703c7f67.png\" height=\"45\" alt=\"LlamaIndex\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcohere\">\u003Cimg src=\"docs\u002Fimages\u002Fexternal\u002Fcohere\u002Fcohere-logo.svg\" height=\"45\" alt=\"Cohere\">\u003C\u002Fa>\n  \u003C\u002Fdiv>\n\u003C\u002Fdiv>\n\n|                                       |                                                               |\n| ------------------------------------- | ------------------------------------------------------------- |\n| 📊 **Replay Analytics and Debugging** | Step-by-step agent execution graphs                           |\n| 💸 **LLM Cost Management**            | Track spend with LLM foundation model providers               |\n| 🤝 **Framework Integrations**         | Native Integrations with CrewAI, AG2 (AutoGen), Agno, LangGraph, & more         |\n| ⚒️ **Self-Host**                      | Want to run AgentOps on your own cloud? You're covered        |\n\n## Quick Start ⌨️\n\n```bash\npip install agentops\n```\n\n\n#### Session replays in 2 lines of code\n\nInitialize the AgentOps client and automatically get analytics on all your LLM calls.\n\n[Get an API key](https:\u002F\u002Fapp.agentops.ai\u002Fsettings\u002Fprojects)\n\n```python\nimport agentops\n\n# Beginning of your program (i.e. main.py, __init__.py)\nagentops.init( \u003C INSERT YOUR API KEY HERE >)\n\n...\n\n# End of program\nagentops.end_session('Success')\n```\n\nAll your sessions can be viewed on the [AgentOps dashboard](https:\u002F\u002Fapp.agentops.ai?ref=gh)\n\u003Cbr\u002F>\n\n## Self-Hosting\n\nLooking to run the full AgentOps app (Dashboard + API backend) on your machine? Follow the setup guide in `app\u002FREADME.md`:\n\n- [Run the App and Backend (Dashboard + API)](app\u002FREADME.md)\n\n\n\u003Cdetails>\n  \u003Csummary>Agent Debugging\u003C\u002Fsummary>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai?ref=gh\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_3f4cdd89ea22.png\" style=\"width: 90%;\" alt=\"Agent Metadata\"\u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai?ref=gh\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_5e73ac81feac.png\" style=\"width: 90%;\" alt=\"Chat Viewer\"\u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai?ref=gh\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_3220651c9c40.png\" style=\"width: 90%;\" alt=\"Event Graphs\"\u002F>\n  \u003C\u002Fa>\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>Session Replays\u003C\u002Fsummary>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai?ref=gh\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_d69ff5c8d756.png\" style=\"width: 90%;\" alt=\"Session Replays\"\u002F>\n  \u003C\u002Fa>\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>Summary Analytics\u003C\u002Fsummary>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai?ref=gh\">\n   \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_ac8f532a4b98.png\" style=\"width: 90%;\" alt=\"Summary Analytics\"\u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai?ref=gh\">\n   \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_30a30ee38661.png\" style=\"width: 90%;\" alt=\"Summary Analytics Charts\"\u002F>\n  \u003C\u002Fa>\n\u003C\u002Fdetails>\n\n\n### First class Developer Experience\nAdd powerful observability to your agents, tools, and functions with as little code as possible: one line at a time.\n\u003Cbr\u002F>\nRefer to our [documentation](http:\u002F\u002Fdocs.agentops.ai)\n\n```python\n# Create a session span (root for all other spans)\nfrom agentops.sdk.decorators import session\n\n@session\ndef my_workflow():\n    # Your session code here\n    return result\n```\n\n```python\n# Create an agent span for tracking agent operations\nfrom agentops.sdk.decorators import agent\n\n@agent\nclass MyAgent:\n    def __init__(self, name):\n        self.name = name\n        \n    # Agent methods here\n```\n\n```python\n# Create operation\u002Ftask spans for tracking specific operations\nfrom agentops.sdk.decorators import operation, task\n\n@operation  # or @task\ndef process_data(data):\n    # Process the data\n    return result\n```\n\n```python\n# Create workflow spans for tracking multi-operation workflows\nfrom agentops.sdk.decorators import workflow\n\n@workflow\ndef my_workflow(data):\n    # Workflow implementation\n    return result\n```\n\n```python\n# Nest decorators for proper span hierarchy\nfrom agentops.sdk.decorators import session, agent, operation\n\n@agent\nclass MyAgent:\n    @operation\n    def nested_operation(self, message):\n        return f\"Processed: {message}\"\n        \n    @operation\n    def main_operation(self):\n        result = self.nested_operation(\"test message\")\n        return result\n\n@session\ndef my_session():\n    agent = MyAgent()\n    return agent.main_operation()\n```\n\nAll decorators support:\n- Input\u002FOutput Recording\n- Exception Handling\n- Async\u002Fawait functions\n- Generator functions\n- Custom attributes and names\n\n## Integrations 🦾\n\n### OpenAI Agents SDK 🖇️\n\nBuild multi-agent systems with tools, handoffs, and guardrails. AgentOps natively integrates with the OpenAI Agents SDKs for both Python and TypeScript.\n\n#### Python\n\n```bash\npip install openai-agents\n```\n\n- [Python integration guide](https:\u002F\u002Fdocs.agentops.ai\u002Fv2\u002Fintegrations\u002Fopenai_agents_python)\n- [OpenAI Agents Python documentation](https:\u002F\u002Fopenai.github.io\u002Fopenai-agents-python\u002F)\n\n#### TypeScript\n\n```bash\nnpm install agentops @openai\u002Fagents\n```\n\n- [TypeScript integration guide](https:\u002F\u002Fdocs.agentops.ai\u002Fv2\u002Fintegrations\u002Fopenai_agents_js)\n- [OpenAI Agents JS documentation](https:\u002F\u002Fopenai.github.io\u002Fopenai-agents-js)\n\n### CrewAI 🛶\n\nBuild Crew agents with observability in just 2 lines of code. Simply set an `AGENTOPS_API_KEY` in your environment, and your crews will get automatic monitoring on the AgentOps dashboard.\n\n```bash\npip install 'crewai[agentops]'\n```\n\n- [AgentOps integration example](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcrewai)\n- [Official CrewAI documentation](https:\u002F\u002Fdocs.crewai.com\u002Fhow-to\u002FAgentOps-Observability)\n\n### AG2 🤖\nWith only two lines of code, add full observability and monitoring to AG2 (formerly AutoGen) agents. Set an `AGENTOPS_API_KEY` in your environment and call `agentops.init()`\n\n- [AG2 Observability Example](https:\u002F\u002Fgithub.com\u002Fag2ai\u002Fag2\u002Fblob\u002Fmain\u002Fnotebook\u002Fagentchat_agentops.ipynb)\n- [AG2 - AgentOps Documentation](https:\u002F\u002Fdocs.ag2.ai\u002Flatest\u002Fdocs\u002Fecosystem\u002Fagentops\u002F)\n\n### Camel AI 🐪\n\nTrack and analyze CAMEL agents with full observability. Set an `AGENTOPS_API_KEY` in your environment and initialize AgentOps to get started.\n\n- [Camel AI](https:\u002F\u002Fwww.camel-ai.org\u002F) - Advanced agent communication framework\n- [AgentOps integration example](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcamel)\n- [Official Camel AI documentation](https:\u002F\u002Fdocs.camel-ai.org\u002Fcookbooks\u002Fagents_tracking.html)\n\n\u003Cdetails>\n  \u003Csummary>Installation\u003C\u002Fsummary>\n\n```bash\npip install \"camel-ai[all]==0.2.11\"\npip install agentops\n```\n\n```python\nimport os\nimport agentops\nfrom camel.agents import ChatAgent\nfrom camel.messages import BaseMessage\nfrom camel.models import ModelFactory\nfrom camel.types import ModelPlatformType, ModelType\n\n# Initialize AgentOps\nagentops.init(os.getenv(\"AGENTOPS_API_KEY\"), tags=[\"CAMEL Example\"])\n\n# Import toolkits after AgentOps init for tracking\nfrom camel.toolkits import SearchToolkit\n\n# Set up the agent with search tools\nsys_msg = BaseMessage.make_assistant_message(\n    role_name='Tools calling operator',\n    content='You are a helpful assistant'\n)\n\n# Configure tools and model\ntools = [*SearchToolkit().get_tools()]\nmodel = ModelFactory.create(\n    model_platform=ModelPlatformType.OPENAI,\n    model_type=ModelType.GPT_4O_MINI,\n)\n\n# Create and run the agent\ncamel_agent = ChatAgent(\n    system_message=sys_msg,\n    model=model,\n    tools=tools,\n)\n\nresponse = camel_agent.step(\"What is AgentOps?\")\nprint(response)\n\nagentops.end_session(\"Success\")\n```\n\nCheck out our [Camel integration guide](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcamel) for more examples including multi-agent scenarios.\n\u003C\u002Fdetails>\n\n### Langchain 🦜🔗\n\nAgentOps works seamlessly with applications built using Langchain. To use the handler, install Langchain as an optional dependency:\n\n\u003Cdetails>\n  \u003Csummary>Installation\u003C\u002Fsummary>\n  \n```shell\npip install agentops[langchain]\n```\n\nTo use the handler, import and set\n\n```python\nimport os\nfrom langchain.chat_models import ChatOpenAI\nfrom langchain.agents import initialize_agent, AgentType\nfrom agentops.integration.callbacks.langchain import LangchainCallbackHandler\n\nAGENTOPS_API_KEY = os.environ['AGENTOPS_API_KEY']\nhandler = LangchainCallbackHandler(api_key=AGENTOPS_API_KEY, tags=['Langchain Example'])\n\nllm = ChatOpenAI(openai_api_key=OPENAI_API_KEY,\n                 callbacks=[handler],\n                 model='gpt-3.5-turbo')\n\nagent = initialize_agent(tools,\n                         llm,\n                         agent=AgentType.CHAT_ZERO_SHOT_REACT_DESCRIPTION,\n                         verbose=True,\n                         callbacks=[handler], # You must pass in a callback handler to record your agent\n                         handle_parsing_errors=True)\n```\n\nCheck out the [Langchain Examples Notebook](.\u002Fexamples\u002Flangchain\u002Flangchain_examples.ipynb) for more details including Async handlers.\n\n\u003C\u002Fdetails>\n\n### Cohere ⌨️\n\nFirst class support for Cohere(>=5.4.0). This is a living integration, should you need any added functionality please message us on Discord!\n\n- [AgentOps integration example](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcohere)\n- [Official Cohere documentation](https:\u002F\u002Fdocs.cohere.com\u002Freference\u002Fabout)\n\n\u003Cdetails>\n  \u003Csummary>Installation\u003C\u002Fsummary>\n  \n```bash\npip install cohere\n```\n\n```python python\nimport cohere\nimport agentops\n\n# Beginning of program's code (i.e. main.py, __init__.py)\nagentops.init(\u003CINSERT YOUR API KEY HERE>)\nco = cohere.Client()\n\nchat = co.chat(\n    message=\"Is it pronounced ceaux-hear or co-hehray?\"\n)\n\nprint(chat)\n\nagentops.end_session('Success')\n```\n\n```python python\nimport cohere\nimport agentops\n\n# Beginning of program's code (i.e. main.py, __init__.py)\nagentops.init(\u003CINSERT YOUR API KEY HERE>)\n\nco = cohere.Client()\n\nstream = co.chat_stream(\n    message=\"Write me a haiku about the synergies between Cohere and AgentOps\"\n)\n\nfor event in stream:\n    if event.event_type == \"text-generation\":\n        print(event.text, end='')\n\nagentops.end_session('Success')\n```\n\u003C\u002Fdetails>\n\n\n### Anthropic ﹨\n\nTrack agents built with the Anthropic Python SDK (>=0.32.0).\n\n- [AgentOps integration guide](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fanthropic)\n- [Official Anthropic documentation](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fwelcome)\n\n\u003Cdetails>\n  \u003Csummary>Installation\u003C\u002Fsummary>\n  \n```bash\npip install anthropic\n```\n\n```python python\nimport anthropic\nimport agentops\n\n# Beginning of program's code (i.e. main.py, __init__.py)\nagentops.init(\u003CINSERT YOUR API KEY HERE>)\n\nclient = anthropic.Anthropic(\n    # This is the default and can be omitted\n    api_key=os.environ.get(\"ANTHROPIC_API_KEY\"),\n)\n\nmessage = client.messages.create(\n        max_tokens=1024,\n        messages=[\n            {\n                \"role\": \"user\",\n                \"content\": \"Tell me a cool fact about AgentOps\",\n            }\n        ],\n        model=\"claude-3-opus-20240229\",\n    )\nprint(message.content)\n\nagentops.end_session('Success')\n```\n\nStreaming\n```python python\nimport anthropic\nimport agentops\n\n# Beginning of program's code (i.e. main.py, __init__.py)\nagentops.init(\u003CINSERT YOUR API KEY HERE>)\n\nclient = anthropic.Anthropic(\n    # This is the default and can be omitted\n    api_key=os.environ.get(\"ANTHROPIC_API_KEY\"),\n)\n\nstream = client.messages.create(\n    max_tokens=1024,\n    model=\"claude-3-opus-20240229\",\n    messages=[\n        {\n            \"role\": \"user\",\n            \"content\": \"Tell me something cool about streaming agents\",\n        }\n    ],\n    stream=True,\n)\n\nresponse = \"\"\nfor event in stream:\n    if event.type == \"content_block_delta\":\n        response += event.delta.text\n    elif event.type == \"message_stop\":\n        print(\"\\n\")\n        print(response)\n        print(\"\\n\")\n```\n\nAsync\n\n```python python\nimport asyncio\nfrom anthropic import AsyncAnthropic\n\nclient = AsyncAnthropic(\n    # This is the default and can be omitted\n    api_key=os.environ.get(\"ANTHROPIC_API_KEY\"),\n)\n\n\nasync def main() -> None:\n    message = await client.messages.create(\n        max_tokens=1024,\n        messages=[\n            {\n                \"role\": \"user\",\n                \"content\": \"Tell me something interesting about async agents\",\n            }\n        ],\n        model=\"claude-3-opus-20240229\",\n    )\n    print(message.content)\n\n\nawait main()\n```\n\u003C\u002Fdetails>\n\n### Mistral 〽️\n\nTrack agents built with the Mistral Python SDK (>=0.32.0).\n\n- [AgentOps integration example](.\u002Fexamples\u002Fmistral\u002F\u002Fmistral_example.ipynb)\n- [Official Mistral documentation](https:\u002F\u002Fdocs.mistral.ai)\n\n\u003Cdetails>\n  \u003Csummary>Installation\u003C\u002Fsummary>\n  \n```bash\npip install mistralai\n```\n\nSync\n\n```python python\nfrom mistralai import Mistral\nimport agentops\n\n# Beginning of program's code (i.e. main.py, __init__.py)\nagentops.init(\u003CINSERT YOUR API KEY HERE>)\n\nclient = Mistral(\n    # This is the default and can be omitted\n    api_key=os.environ.get(\"MISTRAL_API_KEY\"),\n)\n\nmessage = client.chat.complete(\n        messages=[\n            {\n                \"role\": \"user\",\n                \"content\": \"Tell me a cool fact about AgentOps\",\n            }\n        ],\n        model=\"open-mistral-nemo\",\n    )\nprint(message.choices[0].message.content)\n\nagentops.end_session('Success')\n```\n\nStreaming\n\n```python python\nfrom mistralai import Mistral\nimport agentops\n\n# Beginning of program's code (i.e. main.py, __init__.py)\nagentops.init(\u003CINSERT YOUR API KEY HERE>)\n\nclient = Mistral(\n    # This is the default and can be omitted\n    api_key=os.environ.get(\"MISTRAL_API_KEY\"),\n)\n\nmessage = client.chat.stream(\n        messages=[\n            {\n                \"role\": \"user\",\n                \"content\": \"Tell me something cool about streaming agents\",\n            }\n        ],\n        model=\"open-mistral-nemo\",\n    )\n\nresponse = \"\"\nfor event in message:\n    if event.data.choices[0].finish_reason == \"stop\":\n        print(\"\\n\")\n        print(response)\n        print(\"\\n\")\n    else:\n        response += event.text\n\nagentops.end_session('Success')\n```\n\nAsync\n\n```python python\nimport asyncio\nfrom mistralai import Mistral\n\nclient = Mistral(\n    # This is the default and can be omitted\n    api_key=os.environ.get(\"MISTRAL_API_KEY\"),\n)\n\n\nasync def main() -> None:\n    message = await client.chat.complete_async(\n        messages=[\n            {\n                \"role\": \"user\",\n                \"content\": \"Tell me something interesting about async agents\",\n            }\n        ],\n        model=\"open-mistral-nemo\",\n    )\n    print(message.choices[0].message.content)\n\n\nawait main()\n```\n\nAsync Streaming\n\n```python python\nimport asyncio\nfrom mistralai import Mistral\n\nclient = Mistral(\n    # This is the default and can be omitted\n    api_key=os.environ.get(\"MISTRAL_API_KEY\"),\n)\n\n\nasync def main() -> None:\n    message = await client.chat.stream_async(\n        messages=[\n            {\n                \"role\": \"user\",\n                \"content\": \"Tell me something interesting about async streaming agents\",\n            }\n        ],\n        model=\"open-mistral-nemo\",\n    )\n\n    response = \"\"\n    async for event in message:\n        if event.data.choices[0].finish_reason == \"stop\":\n            print(\"\\n\")\n            print(response)\n            print(\"\\n\")\n        else:\n            response += event.text\n\n\nawait main()\n```\n\u003C\u002Fdetails>\n\n\n\n### CamelAI ﹨\n\nTrack agents built with the CamelAI Python SDK (>=0.32.0).\n\n- [CamelAI integration guide](https:\u002F\u002Fdocs.camel-ai.org\u002Fcookbooks\u002Fagents_tracking.html#)\n- [Official CamelAI documentation](https:\u002F\u002Fdocs.camel-ai.org\u002Findex.html)\n\n\u003Cdetails>\n  \u003Csummary>Installation\u003C\u002Fsummary>\n  \n```bash\npip install camel-ai[all]\npip install agentops\n```\n\n```python python\n#Import Dependencies\nimport agentops\nimport os\nfrom getpass import getpass\nfrom dotenv import load_dotenv\n\n#Set Keys\nload_dotenv()\nopenai_api_key = os.getenv(\"OPENAI_API_KEY\") or \"\u003Cyour openai key here>\"\nagentops_api_key = os.getenv(\"AGENTOPS_API_KEY\") or \"\u003Cyour agentops key here>\"\n\n\n\n```\n\u003C\u002Fdetails>\n\n[You can find usage examples here!](examples\u002Fcamelai_examples\u002FREADME.md).\n\n\n\n### LiteLLM 🚅\n\nAgentOps provides support for LiteLLM(>=1.3.1), allowing you to call 100+ LLMs using the same Input\u002FOutput Format. \n\n- [AgentOps integration example](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Flitellm)\n- [Official LiteLLM documentation](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders)\n\n\u003Cdetails>\n  \u003Csummary>Installation\u003C\u002Fsummary>\n  \n```bash\npip install litellm\n```\n\n```python python\n# Do not use LiteLLM like this\n# from litellm import completion\n# ...\n# response = completion(model=\"claude-3\", messages=messages)\n\n# Use LiteLLM like this\nimport litellm\n...\nresponse = litellm.completion(model=\"claude-3\", messages=messages)\n# or\nresponse = await litellm.acompletion(model=\"claude-3\", messages=messages)\n```\n\u003C\u002Fdetails>\n\n### LlamaIndex 🦙\n\n\nAgentOps works seamlessly with applications built using LlamaIndex, a framework for building context-augmented generative AI applications with LLMs.\n\n\u003Cdetails>\n  \u003Csummary>Installation\u003C\u002Fsummary>\n  \n```shell\npip install llama-index-instrumentation-agentops\n```\n\nTo use the handler, import and set\n\n```python\nfrom llama_index.core import set_global_handler\n\n# NOTE: Feel free to set your AgentOps environment variables (e.g., 'AGENTOPS_API_KEY')\n# as outlined in the AgentOps documentation, or pass the equivalent keyword arguments\n# anticipated by AgentOps' AOClient as **eval_params in set_global_handler.\n\nset_global_handler(\"agentops\")\n```\n\nCheck out the [LlamaIndex docs](https:\u002F\u002Fdocs.llamaindex.ai\u002Fen\u002Fstable\u002Fmodule_guides\u002Fobservability\u002F?h=agentops#agentops) for more details.\n\n\u003C\u002Fdetails>\n\n### Llama Stack 🦙🥞\n\nAgentOps provides support for Llama Stack Python Client(>=0.0.53), allowing you to monitor your Agentic applications. \n\n- [AgentOps integration example 1](https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F530\u002Ffiles\u002F65a5ab4fdcf310326f191d4b870d4f553591e3ea#diff-fdddf65549f3714f8f007ce7dfd1cde720329fe54155d54389dd50fbd81813cb)\n- [AgentOps integration example 2](https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F530\u002Ffiles\u002F65a5ab4fdcf310326f191d4b870d4f553591e3ea#diff-6688ff4fb7ab1ce7b1cc9b8362ca27264a3060c16737fb1d850305787a6e3699)\n- [Official Llama Stack Python Client](https:\u002F\u002Fgithub.com\u002Fmeta-llama\u002Fllama-stack-client-python)\n\n### SwarmZero AI 🐝\n\nTrack and analyze SwarmZero agents with full observability. Set an `AGENTOPS_API_KEY` in your environment and initialize AgentOps to get started.\n\n- [SwarmZero](https:\u002F\u002Fswarmzero.ai) - Advanced multi-agent framework\n- [AgentOps integration example](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fswarmzero)\n- [SwarmZero AI integration example](https:\u002F\u002Fdocs.swarmzero.ai\u002Fexamples\u002Fai-agents\u002Fbuild-and-monitor-a-web-search-agent)\n- [SwarmZero AI - AgentOps documentation](https:\u002F\u002Fdocs.swarmzero.ai\u002Fsdk\u002Fobservability\u002Fagentops)\n- [Official SwarmZero Python SDK](https:\u002F\u002Fgithub.com\u002Fswarmzero\u002Fswarmzero)\n\n\u003Cdetails>\n  \u003Csummary>Installation\u003C\u002Fsummary>\n\n```bash\npip install swarmzero\npip install agentops\n```\n\n```python\nfrom dotenv import load_dotenv\nload_dotenv()\n\nimport agentops\nagentops.init(\u003CINSERT YOUR API KEY HERE>)\n\nfrom swarmzero import Agent, Swarm\n# ...\n```\n\u003C\u002Fdetails>\n\n## Evaluations Roadmap 🧭\n\n| Platform                                                                     | Dashboard                                  | Evals                                  |\n| ---------------------------------------------------------------------------- | ------------------------------------------ | -------------------------------------- |\n| ✅ Python SDK                                                                | ✅ Multi-session and Cross-session metrics | ✅ Custom eval metrics                 |\n| 🚧 Evaluation builder API                                                    | ✅ Custom event tag tracking              | 🔜 Agent scorecards                    |\n| 🚧 [Javascript\u002FTypescript SDK (Alpha)](https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops-node) | ✅ Session replays                         | 🔜 Evaluation playground + leaderboard |\n\n## Debugging Roadmap 🧭\n\n| Performance testing                       | Environments                                                                        | LLM Testing                                 | Reasoning and execution testing                   |\n| ----------------------------------------- | ----------------------------------------------------------------------------------- | ------------------------------------------- | ------------------------------------------------- |\n| ✅ Event latency analysis                 | 🔜 Non-stationary environment testing                                               | 🔜 LLM non-deterministic function detection | 🚧 Infinite loops and recursive thought detection |\n| ✅ Agent workflow execution pricing       | 🔜 Multi-modal environments                                                         | 🚧 Token limit overflow flags               | 🔜 Faulty reasoning detection                     |\n| 🚧 Success validators (external)          | 🔜 Execution containers                                                             | 🔜 Context limit overflow flags             | 🔜 Generative code validators                     |\n| 🔜 Agent controllers\u002Fskill tests          | ✅ Honeypot and prompt injection detection ([PromptArmor](https:\u002F\u002Fpromptarmor.com)) | ✅ API bill tracking                        | 🔜 Error breakpoint analysis                      |\n| 🔜 Information context constraint testing | 🔜 Anti-agent roadblocks (i.e. Captchas)                                            | 🔜 CI\u002FCD integration checks                 |                                                   |\n| 🔜 Regression testing                     | ✅ Multi-agent framework visualization                                              |                                             |                                                   |\n\n### Why AgentOps? 🤔\n\nWithout the right tools, AI agents are slow, expensive, and unreliable. Our mission is to bring your agent from prototype to production. Here's why AgentOps stands out:\n\n- **Comprehensive Observability**: Track your AI agents' performance, user interactions, and API usage.\n- **Real-Time Monitoring**: Get instant insights with session replays, metrics, and live monitoring tools.\n- **Cost Control**: Monitor and manage your spend on LLM and API calls.\n- **Failure Detection**: Quickly identify and respond to agent failures and multi-agent interaction issues.\n- **Tool Usage Statistics**: Understand how your agents utilize external tools with detailed analytics.\n- **Session-Wide Metrics**: Gain a holistic view of your agents' sessions with comprehensive statistics.\n\nAgentOps is designed to make agent observability, testing, and monitoring easy.\n\n\n## Star History\n\nCheck out our growth in the community:\n\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_ae28ceeb85d0.png\" style=\"max-width: 500px\" width=\"50%\" alt=\"Logo\">\n\n## Popular projects using AgentOps\n\n\n| Repository | Stars  |\n| :--------  | -----: |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_ca3e395349af.png\" width=\"20\" height=\"20\" alt=\"\">    [geekan](https:\u002F\u002Fgithub.com\u002Fgeekan) \u002F [MetaGPT](https:\u002F\u002Fgithub.com\u002Fgeekan\u002FMetaGPT) | 42787 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_a970f5ee7121.png\" width=\"20\" height=\"20\" alt=\"\">    [run-llama](https:\u002F\u002Fgithub.com\u002Frun-llama) \u002F [llama_index](https:\u002F\u002Fgithub.com\u002Frun-llama\u002Fllama_index) | 34446 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_0da1bc8e33c0.png\" width=\"20\" height=\"20\" alt=\"\">    [crewAIInc](https:\u002F\u002Fgithub.com\u002FcrewAIInc) \u002F [crewAI](https:\u002F\u002Fgithub.com\u002FcrewAIInc\u002FcrewAI) | 18287 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_8d62a422e6bb.png\" width=\"20\" height=\"20\" alt=\"\">    [camel-ai](https:\u002F\u002Fgithub.com\u002Fcamel-ai) \u002F [camel](https:\u002F\u002Fgithub.com\u002Fcamel-ai\u002Fcamel) | 5166 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_f02217d064c9.png\" width=\"20\" height=\"20\" alt=\"\">    [superagent-ai](https:\u002F\u002Fgithub.com\u002Fsuperagent-ai) \u002F [superagent](https:\u002F\u002Fgithub.com\u002Fsuperagent-ai\u002Fsuperagent) | 5050 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_63a53f1b28d4.png\" width=\"20\" height=\"20\" alt=\"\">    [iyaja](https:\u002F\u002Fgithub.com\u002Fiyaja) \u002F [llama-fs](https:\u002F\u002Fgithub.com\u002Fiyaja\u002Fllama-fs) | 4713 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_e2d4ed2ab1f0.png\" width=\"20\" height=\"20\" alt=\"\">    [ag2ai](https:\u002F\u002Fgithub.com\u002Fag2ai) \u002F [ag2](https:\u002F\u002Fgithub.com\u002Fag2ai\u002Fag2) | 4240 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_d714b3dfada0.png\" width=\"20\" height=\"20\" alt=\"\">    [BasedHardware](https:\u002F\u002Fgithub.com\u002FBasedHardware) \u002F [Omi](https:\u002F\u002Fgithub.com\u002FBasedHardware\u002FOmi) | 2723 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_b32659ab3f14.png\" width=\"20\" height=\"20\" alt=\"\">    [MervinPraison](https:\u002F\u002Fgithub.com\u002FMervinPraison) \u002F [PraisonAI](https:\u002F\u002Fgithub.com\u002FMervinPraison\u002FPraisonAI) | 2007 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_71f268c1ce56.png\" width=\"20\" height=\"20\" alt=\"\">    [AgentOps-AI](https:\u002F\u002Fgithub.com\u002FAgentOps-AI) \u002F [Jaiqu](https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002FJaiqu) | 272 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_c830625a2537.png\" width=\"20\" height=\"20\" alt=\"\">    [swarmzero](https:\u002F\u002Fgithub.com\u002Fswarmzero) \u002F [swarmzero](https:\u002F\u002Fgithub.com\u002Fswarmzero\u002Fswarmzero) | 195 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_67bc98429ccd.png\" width=\"20\" height=\"20\" alt=\"\">    [strnad](https:\u002F\u002Fgithub.com\u002Fstrnad) \u002F [CrewAI-Studio](https:\u002F\u002Fgithub.com\u002Fstrnad\u002FCrewAI-Studio) | 134 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_6f262c37134b.png\" width=\"20\" height=\"20\" alt=\"\">    [alejandro-ao](https:\u002F\u002Fgithub.com\u002Falejandro-ao) \u002F [exa-crewai](https:\u002F\u002Fgithub.com\u002Falejandro-ao\u002Fexa-crewai) | 55 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_296184182561.png\" width=\"20\" height=\"20\" alt=\"\">    [tonykipkemboi](https:\u002F\u002Fgithub.com\u002Ftonykipkemboi) \u002F [youtube_yapper_trapper](https:\u002F\u002Fgithub.com\u002Ftonykipkemboi\u002Fyoutube_yapper_trapper) | 47 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_9070806bc48e.png\" width=\"20\" height=\"20\" alt=\"\">    [sethcoast](https:\u002F\u002Fgithub.com\u002Fsethcoast) \u002F [cover-letter-builder](https:\u002F\u002Fgithub.com\u002Fsethcoast\u002Fcover-letter-builder) | 27 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_69c8e77d579d.png\" width=\"20\" height=\"20\" alt=\"\">    [bhancockio](https:\u002F\u002Fgithub.com\u002Fbhancockio) \u002F [chatgpt4o-analysis](https:\u002F\u002Fgithub.com\u002Fbhancockio\u002Fchatgpt4o-analysis) | 19 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_9e8c8e460ac7.png\" width=\"20\" height=\"20\" alt=\"\">    [breakstring](https:\u002F\u002Fgithub.com\u002Fbreakstring) \u002F [Agentic_Story_Book_Workflow](https:\u002F\u002Fgithub.com\u002Fbreakstring\u002FAgentic_Story_Book_Workflow) | 14 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_2785b1bedaed.png\" width=\"20\" height=\"20\" alt=\"\">    [MULTI-ON](https:\u002F\u002Fgithub.com\u002FMULTI-ON) \u002F [multion-python](https:\u002F\u002Fgithub.com\u002FMULTI-ON\u002Fmultion-python) | 13 |\n\n\n_Generated using [github-dependents-info](https:\u002F\u002Fgithub.com\u002Fnvuillam\u002Fgithub-dependents-info), by [Nicolas Vuillamy](https:\u002F\u002Fgithub.com\u002Fnvuillam)_\n","\u003Cdiv align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fagentops.ai?ref=gh\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_1771207e29e0.png\" alt=\"Logo\">\n  \u003C\u002Fa>\n\u003C\u002Fdiv>\n\n\u003Cdiv align=\"center\">\n  \u003Cem>面向AI智能体的可观测性与开发工具平台\u003C\u002Fem>\n\u003C\u002Fdiv>\n\n\u003Cbr \u002F>\n\n\u003Cdiv align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fpepy.tech\u002Fproject\u002Fagentops\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_ae8fe72d2972.png\" alt=\"下载量\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fagentops-ai\u002Fagentops\u002Fissues\">\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fcommit-activity\u002Fm\u002Fagentops-ai\u002Fagentops\" alt=\"git提交活跃度\">\n  \u003C\u002Fa>\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fagentops?&color=3670A0\" alt=\"PyPI - 版本\">\n  \u003Ca href=\"https:\u002F\u002Fopensource.org\u002Flicenses\u002FMIT\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-yellow.svg?&color=3670A0\" alt=\"许可证：MIT\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fsmithery.ai\u002Fserver\u002F@AgentOps-AI\u002Fagentops-mcp\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_2ab22774b7ca.png\"\u002F>\n  \u003C\u002Fa>\n\u003C\u002Fdiv>\n\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Ftwitter.com\u002Fagentopsai\u002F\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Ftwitter\u002Ffollow\u002Fagentopsai?style=social\" alt=\"Twitter\" style=\"height: 20px;\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FFagdcwwXRR\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fdiscord-7289da.svg?style=flat-square&logo=discord\" alt=\"Discord\" style=\"height: 20px;\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai\u002F?ref=gh\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDashboard-blue.svg?style=flat-square\" alt=\"仪表板\" style=\"height: 20px;\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fintroduction\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDocumentation-orange.svg?style=flat-square\" alt=\"文档\" style=\"height: 20px;\">\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fentelligence.ai\u002FAgentOps-AI&agentops\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FChat%20with%20Docs-green.svg?style=flat-square\" alt=\"与文档聊天\" style=\"height: 20px;\">\n  \u003C\u002Fa>\n\u003C\u002Fp>\n\n\u003Cdiv align=\"center\">\n  \u003Cvideo src=\"https:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002Fdfb4fa8d-d8c4-4965-9ff6-5b8514c1c22f\" width=\"650\" autoplay loop muted>\u003C\u002Fvideo>\n\u003C\u002Fdiv>\n\n\u003Cbr\u002F>\n\nAgentOps 帮助开发者构建、评估和监控 AI 智能体，从原型到生产环境。\n\n## 开源\n\nAgentOps 应用程序基于 MIT 许可证开源。您可以在我们的 [应用目录](https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Ftree\u002Fmain\u002Fapp) 中探索代码。\n\n## 关键集成 🔌\n\n\u003Cdiv align=\"center\" style=\"background-color: white; padding: 20px; border-radius: 10px; margin: 0 auto; max-width: 800px;\">\n  \u003Cdiv style=\"display: flex; flex-wrap: wrap; justify-content: center; align-items: center; gap: 30px; margin-bottom: 20px;\">\n    \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fv2\u002Fintegrations\u002Fopenai_agents_python\">\u003Cimg src=\"docs\u002Fimages\u002Fexternal\u002Fopenai\u002Fagents-sdk.svg\" height=\"45\" alt=\"OpenAI Agents SDK\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcrewai\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_359269f184a0.png\" height=\"45\" alt=\"CrewAI\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.ag2.ai\u002Fdocs\u002Fecosystem\u002Fagentops\">\u003Cimg src=\"docs\u002Fimages\u002Fexternal\u002Fag2\u002Fag2-logo.svg\" height=\"45\" alt=\"AG2 (AutoGen)\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fmicrosoft\">\u003Cimg src=\"docs\u002Fimages\u002Fexternal\u002Fmicrosoft\u002Fmicrosoft_logo.svg\" height=\"45\" alt=\"微软\">\u003C\u002Fa>\n  \u003C\u002Fdiv>\n  \n  \u003Cdiv style=\"display: flex; flex-wrap: wrap; justify-content: center; align-items: center; gap: 30px; margin-bottom: 20px;\">\n    \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Flangchain\">\u003Cimg src=\"docs\u002Fimages\u002Fexternal\u002Flangchain\u002Flangchain-logo.svg\" height=\"45\" alt=\"LangChain\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcamel\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_416126dc4ed3.png\" height=\"45\" alt=\"Camel AI\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.llamaindex.ai\u002Fen\u002Fstable\u002Fmodule_guides\u002Fobservability\u002F?h=agentops#agentops\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_2485703c7f67.png\" height=\"45\" alt=\"LlamaIndex\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcohere\">\u003Cimg src=\"docs\u002Fimages\u002Fexternal\u002Fcohere\u002Fcohere-logo.svg\" height=\"45\" alt=\"Cohere\">\u003C\u002Fa>\n  \u003C\u002Fdiv>\n\u003C\u002Fdiv>\n\n|                                       |                                                               |\n| ------------------------------------- | ------------------------------------------------------------- |\n| 📊 **回放分析与调试**                 | 分步骤展示智能体执行图                                      |\n| 💸 **LLM 成本管理**                   | 跟踪与 LLM 基础模型提供商的支出                              |\n| 🤝 **框架集成**                       | 与 CrewAI、AG2（AutoGen）、Agno、LangGraph 等原生集成         |\n| ⚒️ **自托管**                         | 想在自己的云上运行 AgentOps 吗？我们支持                    |\n\n## 快速开始 ⌨️\n\n```bash\npip install agentops\n```\n\n\n#### 两行代码实现会话回放\n\n初始化 AgentOps 客户端，即可自动获取所有 LLM 调用的分析数据。\n\n[获取 API 密钥](https:\u002F\u002Fapp.agentops.ai\u002Fsettings\u002Fprojects)\n\n```python\nimport agentops\n\n# 您程序的开头（如 main.py、__init__.py）\nagentops.init( \u003C 插入您的 API 密钥 >)\n\n...\n\n# 程序结束\nagentops.end_session('成功')\n```\n\n您所有的会话都可以在 [AgentOps 仪表板](https:\u002F\u002Fapp.agentops.ai?ref=gh) 上查看。\n\u003Cbr\u002F>\n\n## 自托管\n\n想在本地运行完整的 AgentOps 应用程序（仪表板 + API 后端）吗？请按照 `app\u002FREADME.md` 中的设置指南操作：\n\n- [运行应用程序和后端（仪表板 + API)](app\u002FREADME.md)\n\n\n\u003Cdetails>\n  \u003Csummary>智能体调试\u003C\u002Fsummary>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai?ref=gh\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_3f4cdd89ea22.png\" style=\"width: 90%;\" alt=\"智能体元数据\"\u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai?ref=gh\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_5e73ac81feac.png\" style=\"width: 90%;\" alt=\"聊天查看器\"\u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai?ref=gh\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_3220651c9c40.png\" style=\"width: 90%;\" alt=\"事件图\"\u002F>\n  \u003C\u002Fa>\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>会话回放\u003C\u002Fsummary>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai?ref=gh\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_d69ff5c8d756.png\" style=\"width: 90%;\" alt=\"会话回放\"\u002F>\n  \u003C\u002Fa>\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>汇总分析\u003C\u002Fsummary>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai?ref=gh\">\n   \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_ac8f532a4b98.png\" style=\"width: 90%;\" alt=\"汇总分析\"\u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fapp.agentops.ai?ref=gh\">\n   \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_30a30ee38661.png\" style=\"width: 90%;\" alt=\"汇总分析图表\"\u002F>\n  \u003C\u002Fa>\n\u003C\u002Fdetails>\n\n\n### 一流的开发者体验\n只需最少的代码——每次一行——即可为您的智能体、工具和函数添加强大的可观测性功能。\n\u003Cbr\u002F>\n请参阅我们的 [文档](http:\u002F\u002Fdocs.agentops.ai)\n\n```python\n\n# 创建会话跨度（作为所有其他跨度的根）\nfrom agentops.sdk.decorators import session\n\n@session\ndef my_workflow():\n    # 您的会话代码在此处\n    return result\n```\n\n```python\n# 创建代理跨度以跟踪代理操作\nfrom agentops.sdk.decorators import agent\n\n@agent\nclass MyAgent:\n    def __init__(self, name):\n        self.name = name\n        \n    # 代理方法在此处\n```\n\n```python\n# 创建操作\u002F任务跨度以跟踪特定操作\nfrom agentops.sdk.decorators import operation, task\n\n@operation  # 或 @task\ndef process_data(data):\n    # 处理数据\n    return result\n```\n\n```python\n# 创建工作流跨度以跟踪多步骤工作流\nfrom agentops.sdk.decorators import workflow\n\n@workflow\ndef my_workflow(data):\n    # 工作流实现\n    return result\n```\n\n```python\n# 嵌套装饰器以构建正确的跨度层级\nfrom agentops.sdk.decorators import session, agent, operation\n\n@agent\nclass MyAgent:\n    @operation\n    def nested_operation(self, message):\n        return f\"Processed: {message}\"\n        \n    @operation\n    def main_operation(self):\n        result = self.nested_operation(\"test message\")\n        return result\n\n@session\ndef my_session():\n    agent = MyAgent()\n    return agent.main_operation()\n```\n\n所有装饰器均支持：\n- 输入\u002F输出记录\n- 异常处理\n- 异步\u002F等待函数\n- 生成器函数\n- 自定义属性和名称\n\n## 集成 🦾\n\n### OpenAI Agents SDK 🖇️\n\n使用工具、交接和护栏构建多代理系统。AgentOps 原生集成 OpenAI 的 Python 和 TypeScript SDK。\n\n#### Python\n\n```bash\npip install openai-agents\n```\n\n- [Python 集成指南](https:\u002F\u002Fdocs.agentops.ai\u002Fv2\u002Fintegrations\u002Fopenai_agents_python)\n- [OpenAI Agents Python 文档](https:\u002F\u002Fopenai.github.io\u002Fopenai-agents-python\u002F)\n\n#### TypeScript\n\n```bash\nnpm install agentops @openai\u002Fagents\n```\n\n- [TypeScript 集成指南](https:\u002F\u002Fdocs.agentops.ai\u002Fv2\u002Fintegrations\u002Fopenai_agents_js)\n- [OpenAI Agents JS 文档](https:\u002F\u002Fopenai.github.io\u002Fopenai-agents-js)\n\n### CrewAI 🛶\n\n只需两行代码即可为 Crew 代理添加可观测性。在您的环境中设置 `AGENTOPS_API_KEY`，您的 Crew 就会在 AgentOps 仪表板上获得自动监控。\n\n```bash\npip install 'crewai[agentops]'\n```\n\n- [AgentOps 集成示例](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcrewai)\n- [CrewAI 官方文档](https:\u002F\u002Fdocs.crewai.com\u002Fhow-to\u002FAgentOps-Observability)\n\n### AG2 🤖\n只需两行代码，即可为 AG2（原 AutoGen）代理添加完整的可观测性和监控功能。在您的环境中设置 `AGENTOPS_API_KEY` 并调用 `agentops.init()`。\n\n- [AG2 可观测性示例](https:\u002F\u002Fgithub.com\u002Fag2ai\u002Fag2\u002Fblob\u002Fmain\u002Fnotebook\u002Fagentchat_agentops.ipynb)\n- [AG2 - AgentOps 文档](https:\u002F\u002Fdocs.ag2.ai\u002Flatest\u002Fdocs\u002Fecosystem\u002Fagentops\u002F)\n\n### Camel AI 🐪\n使用完全可观测性跟踪和分析 CAMEL 代理。在您的环境中设置 `AGENTOPS_API_KEY` 并初始化 AgentOps 即可开始。\n\n- [Camel AI](https:\u002F\u002Fwww.camel-ai.org\u002F) - 先进的代理通信框架\n- [AgentOps 集成示例](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcamel)\n- [官方 Camel AI 文档](https:\u002F\u002Fdocs.camel-ai.org\u002Fcookbooks\u002Fagents_tracking.html)\n\n\u003Cdetails>\n  \u003Csummary>安装\u003C\u002Fsummary>\n\n```bash\npip install \"camel-ai[all]==0.2.11\"\npip install agentops\n```\n\n```python\nimport os\nimport agentops\nfrom camel.agents import ChatAgent\nfrom camel.messages import BaseMessage\nfrom camel.models import ModelFactory\nfrom camel.types import ModelPlatformType, ModelType\n\n# 初始化 AgentOps\nagentops.init(os.getenv(\"AGENTOPS_API_KEY\"), tags=[\"CAMEL 示例\"])\n\n# 在 AgentOps 初始化后导入工具包以便跟踪\nfrom camel.toolkits import SearchToolkit\n\n# 设置带有搜索工具的代理\nsys_msg = BaseMessage.make_assistant_message(\n    role_name='Tools calling operator',\n    content='You are a helpful assistant'\n)\n\n# 配置工具和模型\ntools = [*SearchToolkit().get_tools()]\nmodel = ModelFactory.create(\n    model_platform=ModelPlatformType.OPENAI,\n    model_type=ModelType.GPT_4O_MINI,\n)\n\n# 创建并运行代理\ncamel_agent = ChatAgent(\n    system_message=sys_msg,\n    model=model,\n    tools=tools,\n)\n\nresponse = camel_agent.step(\"What is AgentOps?\")\nprint(response)\n\nagentops.end_session(\"Success\")\n```\n\n请查看我们的 [Camel 集成指南](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcamel)，了解更多包括多代理场景在内的示例。\n\u003C\u002Fdetails>\n\n### Langchain 🦜🔗\nAgentOps 可与基于 Langchain 构建的应用程序无缝协作。要使用该处理器，需将 Langchain 作为可选依赖项安装：\n\n\u003Cdetails>\n  \u003Csummary>安装\u003C\u002Fsummary>\n  \n```shell\npip install agentops[langchain]\n```\n\n要使用该处理器，需导入并设置\n\n```python\nimport os\nfrom langchain.chat_models import ChatOpenAI\nfrom langchain.agents import initialize_agent, AgentType\nfrom agentops.integration.callbacks.langchain import LangchainCallbackHandler\n\nAGENTOPS_API_KEY = os.environ['AGENTOPS_API_KEY']\nhandler = LangchainCallbackHandler(api_key=AGENTOPS_API_KEY, tags=['Langchain 示例'])\n\nllm = ChatOpenAI(openai_api_key=OPENAI_API_KEY,\n                 callbacks=[handler],\n                 model='gpt-3.5-turbo')\n\nagent = initialize_agent(tools,\n                         llm,\n                         agent=AgentType.CHAT_ZERO_SHOT_REACT_DESCRIPTION,\n                         verbose=True,\n                         callbacks=[handler], # 必须传递回调处理器才能记录您的代理\n                         handle_parsing_errors=True)\n```\n\n请参阅 [Langchain 示例笔记本](.\u002Fexamples\u002Flangchain\u002Flangchain_examples.ipynb)，了解更多包括异步处理器在内的详细信息。\n\n\u003C\u002Fdetails>\n\n### Cohere ⌨️\n对 Cohere（>=5.4.0）的一流支持。这是一项持续更新的集成，如果您需要任何额外功能，请在 Discord 上联系我们！\n\n- [AgentOps 集成示例](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fcohere)\n- [Cohere 官方文档](https:\u002F\u002Fdocs.cohere.com\u002Freference\u002Fabout)\n\n\u003Cdetails>\n  \u003Csummary>安装\u003C\u002Fsummary>\n  \n```bash\npip install cohere\n```\n\n```python python\nimport cohere\nimport agentops\n\n# 程序开头代码（如 main.py、__init__.py）\nagentops.init(\u003C插入您的 API 密钥>)\nco = cohere.Client()\n\nchat = co.chat(\n    message=\"Is it pronounced ceaux-hear or co-hehray?\"\n)\n\nprint(chat)\n\nagentops.end_session('Success')\n```\n\n```python python\nimport cohere\nimport agentops\n\n# 程序开头代码（如 main.py、__init__.py）\nagentops.init(\u003C插入您的 API 密钥>)\n\nco = cohere.Client()\n\nstream = co.chat_stream(\n    message=\"Write me a haiku about the synergies between Cohere and AgentOps\"\n)\n\nfor event in stream:\n    if event.event_type == \"text-generation\":\n        print(event.text, end='')\n\nagentops.end_session('Success')\n```\n\u003C\u002Fdetails>\n\n### Anthropic ﹨\n\n跟踪使用 Anthropic Python SDK（>=0.32.0）构建的代理。\n\n- [AgentOps 集成指南](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fanthropic)\n- [Anthropic 官方文档](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fwelcome)\n\n\u003Cdetails>\n  \u003Csummary>安装\u003C\u002Fsummary>\n  \n```bash\npip install anthropic\n```\n\n```python python\nimport anthropic\nimport agentops\n\n# 程序代码的开头（即 main.py、__init__.py）\nagentops.init(\u003C在此插入您的 API 密钥>)\n\nclient = anthropic.Anthropic(\n    # 这是默认值，可以省略\n    api_key=os.environ.get(\"ANTHROPIC_API_KEY\"),\n)\n\nmessage = client.messages.create(\n        max_tokens=1024,\n        messages=[\n            {\n                \"role\": \"user\",\n                \"content\": \"告诉我一个关于 AgentOps 的酷事实\",\n            }\n        ],\n        model=\"claude-3-opus-20240229\",\n    )\nprint(message.content)\n\nagentops.end_session('Success')\n```\n\n流式处理\n```python python\nimport anthropic\nimport agentops\n\n# 程序代码的开头（即 main.py、__init__.py）\nagentops.init(\u003C在此插入您的 API 密钥>)\n\nclient = anthropic.Anthropic(\n    # 这是默认值，可以省略\n    api_key=os.environ.get(\"ANTHROPIC_API_KEY\"),\n)\n\nstream = client.messages.create(\n    max_tokens=1024,\n    model=\"claude-3-opus-20240229\",\n    messages=[\n        {\n            \"role\": \"user\",\n            \"content\": \"跟我讲讲流式代理的一些有趣之处吧\",\n        }\n    ],\n    stream=True,\n)\n\nresponse = \"\"\nfor event in stream:\n    if event.type == \"content_block_delta\":\n        response += event.delta.text\n    elif event.type == \"message_stop\":\n        print(\"\\n\")\n        print(response)\n        print(\"\\n\")\n```\n\n异步\n\n```python python\nimport asyncio\nfrom anthropic import AsyncAnthropic\n\nclient = AsyncAnthropic(\n    # 这是默认值，可以省略\n    api_key=os.environ.get(\"ANTHROPIC_API_KEY\"),\n)\n\n\nasync def main() -> None:\n    message = await client.messages.create(\n        max_tokens=1024,\n        messages=[\n            {\n                \"role\": \"user\",\n                \"content\": \"给我讲讲异步代理的一些有趣之处吧\",\n            }\n        ],\n        model=\"claude-3-opus-20240229\",\n    )\n    print(message.content)\n\n\nawait main()\n```\n\u003C\u002Fdetails>\n\n### Mistral 〽️\n\n跟踪使用 Mistral Python SDK（>=0.32.0）构建的代理。\n\n- [AgentOps 集成示例](.\u002Fexamples\u002Fmistral\u002F\u002Fmistral_example.ipynb)\n- [Mistral 官方文档](https:\u002F\u002Fdocs.mistral.ai)\n\n\u003Cdetails>\n  \u003Csummary>安装\u003C\u002Fsummary>\n  \n```bash\npip install mistralai\n```\n\n同步\n\n```python python\nfrom mistralai import Mistral\nimport agentops\n\n# 程序代码的开头（即 main.py、__init__.py）\nagentops.init(\u003C在此插入您的 API 密钥>)\n\nclient = Mistral(\n    # 这是默认值，可以省略\n    api_key=os.environ.get(\"MISTRAL_API_KEY\"),\n)\n\nmessage = client.chat.complete(\n        messages=[\n            {\n                \"role\": \"user\",\n                \"content\": \"告诉我一个关于 AgentOps 的酷事实\",\n            }\n        ],\n        model=\"open-mistral-nemo\",\n    )\nprint(message.choices[0].message.content)\n\nagentops.end_session('Success')\n```\n\n流式处理\n\n```python python\nfrom mistralai import Mistral\nimport agentops\n\n# 程序代码的开头（即 main.py、__init__.py）\nagentops.init(\u003C在此插入您的 API 密钥>)\n\nclient = Mistral(\n    # 这是默认值，可以省略\n    api_key=os.environ.get(\"MISTRAL_API_KEY\"),\n)\n\nmessage = client.chat.stream(\n        messages=[\n            {\n                \"role\": \"user\",\n                \"content\": \"跟我讲讲流式代理的一些有趣之处吧\",\n            }\n        ],\n        model=\"open-mistral-nemo\",\n    )\n\nresponse = \"\"\nfor event in message:\n    if event.data.choices[0].finish_reason == \"stop\":\n        print(\"\\n\")\n        print(response)\n        print(\"\\n\")\n    else:\n        response += event.text\n\nagentops.end_session('Success')\n```\n\n异步\n\n```python python\nimport asyncio\nfrom mistralai import Mistral\n\nclient = Mistral(\n    # 这是默认值，可以省略\n    api_key=os.environ.get(\"MISTRAL_API_KEY\"),\n)\n\n\nasync def main() -> None:\n    message = await client.chat.complete_async(\n        messages=[\n            {\n                \"role\": \"user\",\n                \"content\": \"跟我讲讲异步代理的一些有趣之处吧\",\n            }\n        ],\n        model=\"open-mistral-nemo\",\n    )\n    print(message.choices[0].message的内容)\n\n\nawait main()\n```\n\n异步流式处理\n\n```python python\nimport asyncio\nfrom mistralai import Mistral\n\nclient = Mistral(\n    # 这是默认值，可以省略\n    api_key=os.environ.get(\"MISTRAL_API_KEY\"),\n)\n\n\nasync def main() -> None:\n    message = await client.chat.stream_async(\n        messages=[\n            {\n                \"role\": \"user\",\n                \"content\": \"跟我讲讲异步流式代理的一些有趣之处吧\",\n            }\n        ],\n        model=\"open-mistral-nemo\",\n    )\n\n    response = \"\"\n    async for event in message:\n        if event.data.choices[0].finish_reason == \"stop\":\n            print(\"\\n\")\n            print(response)\n            print(\"\\n\")\n        else:\n            response += event.text\n\n\nawait main()\n```\n\u003C\u002Fdetails>\n\n\n\n### CamelAI ﹨\n\n跟踪使用 CamelAI Python SDK（>=0.32.0）构建的代理。\n\n- [CamelAI 集成指南](https:\u002F\u002Fdocs.camel-ai.org\u002Fcookbooks\u002Fagents_tracking.html#)\n- [CamelAI 官方文档](https:\u002F\u002Fdocs.camel-ai.org\u002Findex.html)\n\n\u003Cdetails>\n  \u003Csummary>安装\u003C\u002Fsummary>\n  \n```bash\npip install camel-ai[all]\npip install agentops\n```\n\n```python python\n# 导入依赖\nimport agentops\nimport os\nfrom getpass import getpass\nfrom dotenv import load_dotenv\n\n# 设置密钥\nload_dotenv()\nopenai_api_key = os.getenv(\"OPENAI_API_KEY\") 或 \"\u003C您在这里填写 OpenAI 密钥>\"\nagentops_api_key = os.getenv(\"AGENTOPS_API_KEY\") 或 \"\u003C您在这里填写 AgentOps 密钥>\"\n\n\n\n```\n\u003C\u002Fdetails>\n\n[您可以在这里找到使用示例！](examples\u002Fcamelai_examples\u002FREADME.md).\n\n\n\n### LiteLLM 🚅\n\nAgentOps 提供对 LiteLLM（>=1.3.1）的支持，允许您使用相同的输入\u002F输出格式调用 100 多种大语言模型。\n\n- [AgentOps 集成示例](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Flitellm)\n- [LiteLLM 官方文档](https:\u002F\u002Fdocs.litellm.ai\u002Fdocs\u002Fproviders)\n\n\u003Cdetails>\n  \u003Csummary>安装\u003C\u002Fsummary>\n  \n```bash\npip install litellm\n```\n\n```python python\n# 不要这样使用 LiteLLM\n# from litellm import completion\n# ...\n# response = completion(model=\"claude-3\", messages=messages)\n\n# 应该这样使用 LiteLLM\nimport litellm\n...\nresponse = litellm.completion(model=\"claude-3\", messages=messages)\n# 或者\nresponse = await litellm.acompletion(model=\"claude-3\", messages=messages)\n```\n\u003C\u002Fdetails>\n\n### LlamaIndex 🦙\n\n\nAgentOps 可与使用 LlamaIndex 构建的应用程序无缝协作，LlamaIndex 是一个用于借助大语言模型构建上下文增强型生成式 AI 应用程序的框架。\n\n\u003Cdetails>\n  \u003Csummary>安装\u003C\u002Fsummary>\n  \n```shell\npip install llama-index-instrumentation-agentops\n```\n\n要使用该处理器，请导入并设置：\n\n```python\nfrom llama_index.core import set_global_handler\n\n# 注意：您可以根据 AgentOps 文档中的说明设置您的 AgentOps 环境变量（例如 'AGENTOPS_API_KEY'），或者在 set_global_handler 中将 AgentOps 的 AOClient 所需的等效关键字参数作为 **eval_params 传递。\nset_global_handler(\"agentops\")\n```\n\n更多详细信息请参阅 [LlamaIndex 文档](https:\u002F\u002Fdocs.llamaindex.ai\u002Fen\u002Fstable\u002Fmodule_guides\u002Fobservability\u002F?h=agentops#agentops)。\n\n\u003C\u002Fdetails>\n\n### Llama Stack 🦙🥞\n\nAgentOps 提供对 Llama Stack Python 客户端（>=0.0.53）的支持，使您能够监控自己的智能体应用。\n\n- [AgentOps 集成示例 1](https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F530\u002Ffiles\u002F65a5ab4fdcf310326f191d4b870d4f553591e3ea#diff-fdddf65549f3714f8f007ce7dfd1cde720329fe54155d54389dd50fbd81813cb)\n- [AgentOps 集成示例 2](https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F530\u002Ffiles\u002F65a5ab4fdcf310326f191d4b870d4f553591e3ea#diff-6688ff4fb7ab1ce7b1cc9b8362ca27264a3060c16737fb1d850305787a6e3699)\n- [官方 Llama Stack Python 客户端](https:\u002F\u002Fgithub.com\u002Fmeta-llama\u002Fllama-stack-client-python)\n\n### SwarmZero AI 🐝\n\n通过全面的可观测性跟踪和分析 SwarmZero 智能体。在您的环境中设置 `AGENTOPS_API_KEY` 并初始化 AgentOps 即可开始使用。\n\n- [SwarmZero](https:\u002F\u002Fswarmzero.ai) - 高级多智能体框架\n- [AgentOps 集成示例](https:\u002F\u002Fdocs.agentops.ai\u002Fv1\u002Fintegrations\u002Fswarmzero)\n- [SwarmZero AI 集成示例](https:\u002F\u002Fdocs.swarmzero.ai\u002Fexamples\u002Fai-agents\u002Fbuild-and-monitor-a-web-search-agent)\n- [SwarmZero AI - AgentOps 文档](https:\u002F\u002Fdocs.swarmzero.ai\u002Fsdk\u002Fobservability\u002Fagentops)\n- [官方 SwarmZero Python SDK](https:\u002F\u002Fgithub.com\u002Fswarmzero\u002Fswarmzero)\n\n\u003Cdetails>\n  \u003Csummary>安装\u003C\u002Fsummary>\n\n```bash\npip install swarmzero\npip install agentops\n```\n\n```python\nfrom dotenv import load_dotenv\nload_dotenv()\n\nimport agentops\nagentops.init(\u003C在此处插入您的 API 密钥>)\n\nfrom swarmzero import Agent, Swarm\n# ...\n```\n\u003C\u002Fdetails>\n\n## 评估路线图 🧭\n\n| 平台                                                                     | 仪表板                                  | 评估                                  |\n| ---------------------------------------------------------------------------- | ------------------------------------------ | -------------------------------------- |\n| ✅ Python SDK                                                                | ✅ 多会话及跨会话指标 | ✅ 自定义评估指标                 |\n| 🚧 评估构建器 API                                                    | ✅ 自定义事件标签追踪              | 🔜 智能体评分卡                    |\n| 🚧 [Javascript\u002FTypescript SDK (Alpha)](https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops-node) | ✅ 会话回放                         | 🔜 评估游乐场 + 排行榜 |\n\n## 调试路线图 🧭\n\n| 性能测试                       | 环境                                                                        | LLM 测试                                 | 推理与执行测试                   |\n| ----------------------------------------- | ----------------------------------------------------------------------------------- | ------------------------------------------- | ------------------------------------------------- |\n| ✅ 事件延迟分析                 | 🔜 非平稳环境测试                                               | 🔜 LLM 非确定性函数检测 | 🚧 无限循环和递归思维检测 |\n| ✅ 智能体工作流执行定价       | 🔜 多模态环境                                                         | 🚧 令牌限制溢出标志               | 🔜 错误推理检测                     |\n| 🚧 成功验证器（外部）          | 🔜 执行容器                                                             | 🔜 上下文限制溢出标志             | 🔜 生成式代码验证器                     |\n| 🔜 智能体控制器\u002F技能测试          | ✅ 蜜罐和提示注入检测 ([PromptArmor](https:\u002F\u002Fpromptarmor.com)) | ✅ API 账单追踪                        | 🔜 错误断点分析                      |\n| 🔜 信息上下文约束测试             | 🔜 反智能体障碍（即验证码）                                            | 🔜 CI\u002FCD 集成检查                 |                                                   |\n| 🔜 回归测试                     | ✅ 多智能体框架可视化                                              |                                             |                                                   |\n\n### 为什么选择 AgentOps？🤔\n\n如果没有合适的工具，AI 智能体将会运行缓慢、成本高昂且不可靠。我们的使命是帮助您将智能体从原型阶段推向生产环境。以下是 AgentOps 的突出优势：\n\n- **全面的可观测性**：跟踪您的 AI 智能体性能、用户交互以及 API 使用情况。\n- **实时监控**：通过会话回放、指标和实时监控工具，即时获取洞察。\n- **成本控制**：监控并管理您在 LLM 和 API 调用上的支出。\n- **故障检测**：快速识别并响应智能体故障及多智能体交互问题。\n- **工具使用统计**：通过详细分析了解您的智能体如何利用外部工具。\n- **会话级指标**：通过全面的统计数据，获得对智能体会话的整体视图。\n\nAgentOps 旨在让智能体的可观测性、测试和监控变得简单易行。\n\n\n## 星标历史\n\n查看我们在社区中的成长历程：\n\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_ae28ceeb85d0.png\" style=\"max-width: 500px\" width=\"50%\" alt=\"Logo\">\n\n## 使用 AgentOps 的热门项目\n\n\n| 仓库 | 星标  |\n| :--------  | -----: |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_ca3e395349af.png\" width=\"20\" height=\"20\" alt=\"\">    [geekan](https:\u002F\u002Fgithub.com\u002Fgeekan) \u002F [MetaGPT](https:\u002F\u002Fgithub.com\u002Fgeekan\u002FMetaGPT) | 42787 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_a970f5ee7121.png\" width=\"20\" height=\"20\" alt=\"\">    [run-llama](https:\u002F\u002Fgithub.com\u002Frun-llama) \u002F [llama_index](https:\u002F\u002Fgithub.com\u002Frun-llama\u002Fllama_index) | 34446 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_0da1bc8e33c0.png\" width=\"20\" height=\"20\" alt=\"\">    [crewAIInc](https:\u002F\u002Fgithub.com\u002FcrewAIInc) \u002F [crewAI](https:\u002F\u002Fgithub.com\u002FcrewAIInc\u002FcrewAI) | 18287 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_8d62a422e6bb.png\" width=\"20\" height=\"20\" alt=\"\">    [camel-ai](https:\u002F\u002Fgithub.com\u002Fcamel-ai) \u002F [camel](https:\u002F\u002Fgithub.com\u002Fcamel-ai\u002Fcamel) | 5166 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_f02217d064c9.png\" width=\"20\" height=\"20\" alt=\"\">    [superagent-ai](https:\u002F\u002Fgithub.com\u002Fsuperagent-ai) \u002F [superagent](https:\u002F\u002Fgithub.com\u002Fsuperagent-ai\u002Fsuperagent) | 5050 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_63a53f1b28d4.png\" width=\"20\" height=\"20\" alt=\"\">    [iyaja](https:\u002F\u002Fgithub.com\u002Fiyaja) \u002F [llama-fs](https:\u002F\u002Fgithub.com\u002Fiyaja\u002Fllama-fs) | 4713 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_e2d4ed2ab1f0.png\" width=\"20\" height=\"20\" alt=\"\">    [ag2ai](https:\u002F\u002Fgithub.com\u002Fag2ai) \u002F [ag2](https:\u002F\u002Fgithub.com\u002Fag2ai\u002Fag2) | 4240 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_d714b3dfada0.png\" width=\"20\" height=\"20\" alt=\"\">    [BasedHardware](https:\u002F\u002Fgithub.com\u002FBasedHardware) \u002F [Omi](https:\u002F\u002Fgithub.com\u002FBasedHardware\u002FOmi) | 2723 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_b32659ab3f14.png\" width=\"20\" height=\"20\" alt=\"\">    [MervinPraison](https:\u002F\u002Fgithub.com\u002FMervinPraison) \u002F [PraisonAI](https:\u002F\u002Fgithub.com\u002FMervinPraison\u002FPraisonAI) | 2007 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_71f268c1ce56.png\" width=\"20\" height=\"20\" alt=\"\">    [AgentOps-AI](https:\u002F\u002Fgithub.com\u002FAgentOps-AI) \u002F [Jaiqu](https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002FJaiqu) | 272 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_c830625a2537.png\" width=\"20\" height=\"20\" alt=\"\">    [swarmzero](https:\u002F\u002Fgithub.com\u002Fswarmzero) \u002F [swarmzero](https:\u002F\u002Fgithub.com\u002Fswarmzero\u002Fswarmzero) | 195 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_67bc98429ccd.png\" width=\"20\" height=\"20\" alt=\"\">    [strnad](https:\u002F\u002Fgithub.com\u002Fstrnad) \u002F [CrewAI-Studio](https:\u002F\u002Fgithub.com\u002Fstrnad\u002FCrewAI-Studio) | 134 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_6f262c37134b.png\" width=\"20\" height=\"20\" alt=\"\">    [alejandro-ao](https:\u002F\u002Fgithub.com\u002Falejandro-ao) \u002F [exa-crewai](https:\u002F\u002Fgithub.com\u002Falejandro-ao\u002Fexa-crewai) | 55 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_296184182561.png\" width=\"20\" height=\"20\" alt=\"\">    [tonykipkemboi](https:\u002F\u002Fgithub.com\u002Ftonykipkemboi) \u002F [youtube_yapper_trapper](https:\u002F\u002Fgithub.com\u002Ftonykipkemboi\u002Fyoutube_yapper_trapper) | 47 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_9070806bc48e.png\" width=\"20\" height=\"20\" alt=\"\">    [sethcoast](https:\u002F\u002Fgithub.com\u002Fsethcoast) \u002F [cover-letter-builder](https:\u002F\u002Fgithub.com\u002Fsethcoast\u002Fcover-letter-builder) | 27 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_69c8e77d579d.png\" width=\"20\" height=\"20\" alt=\"\">    [bhancockio](https:\u002F\u002Fgithub.com\u002Fbhancockio) \u002F [chatgpt4o-analysis](https:\u002F\u002Fgithub.com\u002Fbhancockio\u002Fchatgpt4o-analysis) | 19 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_9e8c8e460ac7.png\" width=\"20\" height=\"20\" alt=\"\">    [breakstring](https:\u002F\u002Fgithub.com\u002Fbreakstring) \u002F [Agentic_Story_Book_Workflow](https:\u002F\u002Fgithub.com\u002Fbreakstring\u002FAgentic_Story_Book_Workflow) | 14 |\n|\u003Cimg class=\"avatar mr-2\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_readme_2785b1bedaed.png\" width=\"20\" height=\"20\" alt=\"\">    [MULTI-ON](https:\u002F\u002Fgithub.com\u002FMULTI-ON) \u002F [multion-python](https:\u002F\u002Fgithub.com\u002FMULTI-ON\u002Fmultion-python) | 13 |\n\n\n_由 [Nicolas Vuillamy](https:\u002F\u002Fgithub.com\u002Fnvuillam) 使用 [github-dependents-info](https:\u002F\u002Fgithub.com\u002Fnvuillam\u002Fgithub-dependents-info) 生成_","# AgentOps 快速上手指南\n\nAgentOps 是一个专为 AI Agent 设计的可观测性与开发工具平台，帮助开发者从原型到生产阶段构建、评估和监控 AI 智能体。\n\n## 环境准备\n\n- **系统要求**：支持 Windows、macOS、Linux\n- **Python 版本**：Python 3.8+\n- **前置依赖**：无特殊系统依赖，需具备 Python 包管理工具 (pip)\n- **API Key**：使用前需在 [AgentOps Dashboard](https:\u002F\u002Fapp.agentops.ai\u002Fsettings\u002Fprojects) 注册并获取 API Key\n\n## 安装步骤\n\n通过 pip 安装核心库：\n\n```bash\npip install agentops\n```\n\n> **提示**：国内开发者若遇到下载缓慢，可指定清华或阿里镜像源加速：\n> ```bash\n> pip install agentops -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n> ```\n\n若需使用特定框架集成（如 LangChain），可安装可选依赖：\n\n```bash\n# 示例：安装 LangChain 集成支持\npip install 'agentops[langchain]'\n```\n\n## 基本使用\n\n只需两行代码即可开启会话监控与回放功能。\n\n### 1. 初始化与结束会话\n\n在你的程序入口文件（如 `main.py`）中添加以下代码：\n\n```python\nimport agentops\n\n# 程序开始时初始化（填入你的 API Key）\nagentops.init(\"\u003CINSERT YOUR API KEY HERE>\")\n\n# ... 你的 AI Agent 业务逻辑 ...\n\n# 程序结束时标记会话状态\nagentops.end_session('Success')\n```\n\n所有会话数据将自动同步至 [AgentOps Dashboard](https:\u002F\u002Fapp.agentops.ai)，你可在此查看执行图谱、LLM 成本及调试信息。\n\n### 2. 高级用法：装饰器追踪\n\nAgentOps 提供原生装饰器，可精细化追踪工作流、智能体及具体操作：\n\n```python\nfrom agentops.sdk.decorators import session, agent, operation\n\n# 定义会话根节点\n@session\ndef my_workflow():\n    # 实例化智能体\n    my_agent = MyAgent()\n    return my_agent.run()\n\n# 定义智能体类\n@agent\nclass MyAgent:\n    def __init__(self):\n        pass\n        \n    # 定义具体操作\n    @operation\n    def run(self):\n        # 执行业务逻辑\n        return \"Task Completed\"\n\n# 执行工作流\nif __name__ == \"__main__\":\n    agentops.init(\"\u003CINSERT YOUR API KEY HERE>\")\n    my_workflow()\n    agentops.end_session('Success')\n```\n\n**特性支持**：\n- 自动记录输入\u002F输出\n- 异常捕获处理\n- 支持 async\u002Fawait 异步函数\n- 支持生成器函数\n- 支持自定义属性与命名","某电商初创团队正在开发一个基于 CrewAI 的多智能体客服系统，旨在自动处理退货申请、查询订单状态及推荐替代商品。\n\n### 没有 agentops 时\n- **黑盒运行难排查**：当智能体错误拒绝合法退货请求时，开发者无法回溯是哪个环节的智能体做出了错误判断，只能盲目打印大量日志。\n- **成本失控无感知**：由于缺乏细粒度监控，团队直到月底收到巨额 API 账单时，才发现某个测试循环在后台疯狂调用高价模型。\n- **效果评估靠直觉**：优化提示词（Prompt）后，无法量化对比新旧版本的回复准确率，迭代升级全靠人工抽检和主观猜测。\n- **生产环境裸奔**：部署上线后，若智能体陷入死循环或响应超时，团队往往先接到用户投诉才知道系统已故障。\n\n### 使用 agentops 后\n- **全链路可视化追踪**：通过 agentops 的仪表盘，开发者能清晰看到每个任务的生命周期，迅速定位到是“政策审核智能体”误读了退货条款。\n- **实时成本与性能监控**：agentops 实时统计每次调用的 Token 消耗和延迟，帮助团队立即发现并修复了导致费用激增的代码逻辑。\n- **数据驱动的基准测试**：利用内置的 Benchmarking 功能，团队可直观对比不同提示词版本的成功率，用数据指导模型优化方向。\n- **主动式异常告警**：设置阈值后，一旦智能体执行失败率飙升或耗时过长，agentops 会立即通知开发人员，在用户感知前解决问题。\n\nagentops 将 AI 智能体的开发从“盲人摸象”转变为可观测、可度量、可优化的工程化流程，显著降低了从原型到生产的落地风险。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FAgentOps-AI_agentops_1771207e.png","AgentOps-AI","AgentOps","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002FAgentOps-AI_4ecf0b8a.png","Developer platform to test and debug AI agents.",null,"AgentOpsAI","agentops.ai","https:\u002F\u002Fgithub.com\u002FAgentOps-AI",[24,28,32,36,40,44,48,51,55,57],{"name":25,"color":26,"percentage":27},"Python","#3572A5",60,{"name":29,"color":30,"percentage":31},"TypeScript","#3178c6",37.5,{"name":33,"color":34,"percentage":35},"PLpgSQL","#336790",0.9,{"name":37,"color":38,"percentage":39},"Shell","#89e051",0.6,{"name":41,"color":42,"percentage":43},"CSS","#663399",0.4,{"name":45,"color":46,"percentage":47},"JavaScript","#f1e05a",0.2,{"name":49,"color":50,"percentage":47},"Jinja","#a52a22",{"name":52,"color":53,"percentage":54},"Dockerfile","#384d54",0.1,{"name":56,"color":53,"percentage":54},"Just",{"name":58,"color":59,"percentage":60},"HTML","#e34c26",0,5429,557,"2026-04-03T12:06:35","MIT",2,"未说明",{"notes":68,"python":66,"dependencies":69},"该工具主要为 Python 库及可自托管的 Dashboard 后端。作为库使用时，仅需通过 pip 安装并配置 API Key 即可，无特殊硬件要求。若选择自托管完整应用（Dashboard + API 后端），需参考 app\u002FREADME.md 进行部署。支持多种 AI 框架集成（如 CrewAI, LangChain, OpenAI Agents SDK 等）。",[70,71,72,73,74],"crewai","openai-agents","langchain","camel-ai","ag2 (autogen)",[76,77,78,79,80],"其他","图像","开发框架","Agent","语言模型",[82,6,83,84,85,86,87,88,89,70,90,72,91,92,93,94,71],"agent","ai","evals","evaluation-metrics","llm","anthropic","autogen","cost-estimation","groq","mistral","ollama","openai","agents-sdk","ready","2026-03-27T02:49:30.150509","2026-04-06T06:55:39.831794",[99,104,109,114,119,123],{"id":100,"question_zh":101,"answer_zh":102,"source_url":103},13108,"为什么导入 agentops 时会报错提示找不到 'anthropic' 模块？","这是因为早期的 instrumentation 初始化逻辑默认加载了所有 LLM 提供商的依赖。维护者已确认这是一个问题，并计划通过补丁修复，目标是让 AgentOps 不再强制要求安装非必要的 LLM 提供商库（如 anthropic）。如果遇到此问题，请确保升级到最新版本，或者暂时安装缺失的依赖包作为临时解决方案。","https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fissues\u002F761",{"id":105,"question_zh":106,"answer_zh":107,"source_url":108},13109,"在 Jupyter Notebook (.ipynb) 中运行 AgentOps 代码时无法初始化，但在 .py 文件中正常，如何解决？","这通常是由于旧版本中的 Bug 导致的。维护者建议在终端运行以下命令升级 AgentOps 到最新版本以获取最新的 Bug 修复：\npip install -U agentops\n升级后请重启 Notebook 内核再次尝试。如果问题仍然存在，可能是特定环境配置问题，建议检查是否完整加载了环境变量。","https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fissues\u002F875",{"id":110,"question_zh":111,"answer_zh":112,"source_url":113},13110,"如何在文档或 README 中找到使用 Anthropic SDK 的示例代码？","官方文档和 README 正在逐步完善对 Anthropic SDK 的支持。目前社区贡献者已提交了包含示例代码片段和 Notebook 示例的 Pull Request。建议查看项目最新的文档页面或 README 文件，其中应包含针对 Anthropic 的集成示例。如果需要更详细的指南，可以关注官方发布的 AI 开发指南，其中会解释上下文、Token 和模型的工作原理。","https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fissues\u002F446",{"id":115,"question_zh":116,"answer_zh":117,"source_url":118},13111,"在 Session Replay 图表上悬停鼠标时出现\"Application error: a client-side exception\"错误怎么办？","这是一个前端显示 Bug，通常表现为 TypeError: e.substring is not a function 或数据显示为 [object Object]。该问题已在后续的仪表板更新中修复。如果遇到此问题，请尝试刷新页面或清除浏览器缓存。如果问题依旧，请确认您使用的是最新版本的 Dashboard，因为后端团队已修复了动作输出显示不正确的问题。","https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fissues\u002F151",{"id":120,"question_zh":121,"answer_zh":122,"source_url":118},13112,"使用 AgentOps 时，如何减少日志中的样板信息并只查看相关内容？","您可以使用自定义的 Callback Handler 来获得更多控制权。虽然官方的 langchain_callback_handler 已经修复了基本显示问题，但自定义 Handler 允许您过滤掉不必要的样板日志，仅打印相关的关键内容。此外，随着 CrewAI 集成的改进（相关 PR 已合并），未来的版本将更容易实现这种精细化控制，无需编写复杂的集成代码。",{"id":124,"question_zh":125,"answer_zh":126,"source_url":127},13113,"在 app.agentops.ai 注册或登录时遇到错误提示怎么办？","部分用户曾反馈在注册或重置密码后无法登录的问题。维护者已发布补丁修复了欢迎页面显示和密码重置链接的问题。如果您仍遇到\"凭证错误\"的提示，请尝试以下步骤：1. 确认邮箱验证是否成功完成；2. 尝试完全退出并重新登录；3. 如果问题持续，请联系支持团队并提供您的用户 ID 以便调试。大多数此类问题已通过最近的更新解决。","https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fissues\u002F557",[129,134,139,144,149,154,159,164,169,174,179,184,189,194,199,204,209,214,219,224],{"id":130,"version":131,"summary_zh":132,"released_at":133},71778,"0.4.21","## 变更内容\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1222 中重新添加了库\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1223 中更新了 .gitignore 文件\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1224 中新增了截图\n* 功能（同步）：从 AgentOps.Next 同步 OSS 分发（$(cd ..\u002Fnext-repo && git rev-parse --short HEAD)），由 @devin-ai-integration[bot] 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1226 中完成\n* 功能（同步）：从 AgentOps.Next 同步文档更新，由 @devin-ai-integration[bot] 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1227 中完成\n* 文档（应用）：添加初学者快速入门和本地运行全栈应用的故障排除指南，由 @devin-ai-integration[bot] 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1228 中完成\n* 添加可运行的 Haystack 示例、集成到 CI 并更新文档，由 @devin-ai-integration[bot] 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1230 中完成\n* 功能（Haystack）：自动监控 Haystack 2.x 的生成器，包括 AzureOpenAIChatGenerator；添加 Azure 示例及文档，由 @devin-ai-integration[bot] 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1231 中完成\n* 更新 pyproject.toml 文件，由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1232 中完成\n\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.20...0.4.21","2025-08-29T06:36:28",{"id":135,"version":136,"summary_zh":137,"released_at":138},71779,"0.4.20","## 变更内容\n* 由 @srilaasya 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1189 中为文档添加视频嵌入\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1190 中发布 AgentOps 开源版本\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1194 中更新文档中的后端运行说明\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1196 中更新 README.md\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1197 中更新 README.md\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1201 中更新 README.md\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1208 中移除未计划的功能\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1206 中修复使用 AgentOps 时的 SSE 流式函数调用错误\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1211 中进行优化\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1212 中使用 Ruff 格式化代码\n\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.19...0.4.20","2025-08-15T16:40:36",{"id":140,"version":141,"summary_zh":142,"released_at":143},71780,"0.4.19","## 变更内容\n* 文档：由 @github-actions[bot] 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1154 中更新了笔记本中的示例\n* 创建 OpenAI 响应 API 测试，由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1156 中完成\n* 从 pyautogen 迁移到 ag2 库，由 @AG2AI-Admin 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1160 中完成\n* 添加 MseeP.ai 徽章，由 @lwsinclair 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1158 中完成\n* 更新 LangChain 笔记本，由 @emmanuel-ferdman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1166 中完成\n* 更新 README.md，由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1168 中完成\n* 为 AI 模型文档添加 llms.txt 编译系统，由 @devin-ai-integration[bot] 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1179 中完成\n* 增强 AG2 仪器化功能，由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1146 中完成\n* 重构 Agno token 指标，由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1164 中完成\n* 修复 Agno 工具集成笔记本，由 @dot-agi 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1167 中完成\n* Xpander 仪器化，由 @srilaasya 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1163 中完成\n* DSPy 回调处理器，由 @michi-okahata 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1161 中完成\n* 增强 AgentOps SDK 的异步功能和日志记录，由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1162 中完成\n* 更新 instrument_logging.py，由 @HarikrishnanK9 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1150 中完成\n* 更新 pyproject.toml，由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1185 中完成\n\n## 新贡献者\n* @AG2AI-Admin 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1160 中完成了首次贡献\n* @lwsinclair 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1158 中完成了首次贡献\n* @HarikrishnanK9 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1150 中完成了首次贡献\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.18...0.4.19","2025-08-01T04:40:52",{"id":145,"version":146,"summary_zh":147,"released_at":148},71781,"0.4.18","## 变更内容\n* [测试] 由 @bboynton97 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1126 中为 dopamine 增加了覆盖率测试\n* 由 @michi-okahata 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1129 中修复了 MCP 文档\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1134 中改进了 Cursor\u002FAgno 工具集成的 Python 输出 d88d\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1135 中添加了示例集成测试\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1137 中修复了引号问题\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1139 中简化了文档侧边栏并修复了图片\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1142 中更新了 introduction.mdx 文件\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1143 中在文档中新增了 Mintlify 的 MCP 文档部分\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1144 中调整了 LLM 提供商和框架章节的顺序\n* 文档：由 @github-actions[bot] 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1138 中更新了基于笔记本的示例\n* 由 @ted-tai 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1147 中修复了使用 MCPToolset 对代理进行插桩时可能出现的错误\n* 由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1152 中在 V3Client 认证响应中添加了项目高级状态检查，并对免费套餐发出警告\n* 由 @dot-agi 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1153 中修复了 OpenAI Chat Completions 工具调用问题\n* 由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1148 中新增了用于 HTTP 跟踪的 @track_endpoint 装饰器\n* 由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1145 中修复了示例代码\n* 由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1155 中将版本号升级至 0.4.18\n\n## 新贡献者\n* @ted-tai 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1147 中完成了首次贡献\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.17...0.4.18","2025-07-17T00:45:55",{"id":150,"version":151,"summary_zh":152,"released_at":153},71782,"0.4.17","## 变更内容\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1080 中添加了 OpenAI Agents SDK 的独立文档\n* 由 @devin-ai-integration 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1097 中更新了会话管理文档，以反映当前基于追踪的架构\n* MCP，由 @michi-okahata 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1108 中实现\n* 删除 MCP，由 @michi-okahata 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1111 中完成\n* 文档：由 @github-actions 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1114 中更新了来自笔记本的示例\n* Smithery MCP 徽章，由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1115 中添加\n* agno 文档中的示例链接损坏，由 @srilaasya 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1116 中修复\n* MCP 文档，由 @michi-okahata 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1118 中编写\n* 限制跨度资源属性，由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1120 中实现\n* Agno 增强功能，由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1112 中完成\n* 文档：由 @github-actions 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1121 中更新了来自笔记本的示例\n* 并发修复，由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1113 中完成\n* 功能：添加 update_trace_metadata 函数，用于动态追踪元数据，由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1100 中实现\n* 添加 LangGraph 与 AgentOps 的集成，由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1107 中完成\n* 修复插桩问题，由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1123 中完成\n* 将 pyproject.toml 中的版本号升级至 0.4.17，由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1124 中完成\n\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.16...0.4.17","2025-07-01T19:43:09",{"id":155,"version":156,"summary_zh":157,"released_at":158},71783,"0.4.16","## 变更内容\n* 修复：@HowieG 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1087 中修复了 README 中损坏的 AG2 图片。\n* 重构：@Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1085 中重构了监控模块，以提升结构和一致性。\n* 文档：@github-actions 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1090 中更新了来自笔记本的示例。\n* 修复：@the-praxs 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1089 中与最新的示例笔记本同步。\n* 新功能：@fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1091 中添加了 OpenAI 流式传输支持，并集成遥测监控。\n* 重构：@fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1092 中简化了 AgnoInstrumentor 中的方法包装，以避免循环…。\n* 更新至 0.4.16：@Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1093 中完成。\n\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.15...0.4.16","2025-06-19T00:51:33",{"id":160,"version":161,"summary_zh":162,"released_at":163},71784,"0.4.15","## 变更内容\n* 修复：通过 @the-praxs 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1031 中的提交，修复了 Linting 问题，使其通过静态检查。\n* 添加 Google ADK 示例 Notebook，用于人工审批工作流，由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1020 中完成。\n* 修复 psutil 冲突，由 @michi-okahata 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1042 中完成。\n* 进行了一些耦合性改进，由 @tcdent 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1030 中完成。\n* 增强 ADK 功能，由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1034 中完成。\n* 新增功能：支持使用 `with` 语句自动管理追踪记录的上下文管理器，由 @the-praxs 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1019 中实现。\n* 重构：更新了插桩管理器的使用方式，并改进了追踪提供者的访问权限，由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1054 中完成。\n* 更新文档，提及处于 Alpha 阶段的 TypeScript SDK，由 @devin-ai-integration 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1053 中完成。\n* 清理 v2 文档链接，由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1058 中完成。\n* 修复：更新了示例以及文档页面中的相关集成和示例，由 @the-praxs 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1029 中完成。\n* 修复：将 GitHub Actions 生产化，以添加 Notebook 文件，由 @the-praxs 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1059 中完成。\n* 从 SDK 中移除第三方模式，由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1047 中完成。\n* 修复文档环境配置参数，由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1060 中完成。\n* 输入输出护栏装饰器，由 @michi-okahata 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1039 中完成。\n* Jupyter Notebook 中的终端日志功能，由 @michi-okahata 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1049 中完成。\n* 新增功能：对代码库进行清理，以提高可维护性和可读性，由 @the-praxs 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1055 中完成。\n* 修复 xAI 图标路径，由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1065 中完成。\n* 修复 CrewAI LLM 相关文档，由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1067 中完成。\n* Agno 插桩功能，由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1050 中完成。\n* 实现 OpenTelemetry 上下文的 concurrent.futures 插桩功能，由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1018 中完成。\n* Smolagents 插桩功能，由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1070 中完成。\n* 文档：更新 Notebook 中的示例，由 @github-actions 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1071 中完成。\n* 在 OpenAI Agents 的 Notebook 中静默执行 pip 安装，由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1072 中完成。\n* 文档：更新 Notebook 中的示例，由 @github-actions 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1073 中完成。\n* 添加 Mem0 插桩功能，由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1027 中完成。\n* 文档新增功能：添加 LlamaIndex 集成文档，由 @devin-ai-integration 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1064 中完成。\n* 文档：更新来自 no 的示例","2025-06-17T00:00:07",{"id":165,"version":166,"summary_zh":167,"released_at":168},71785,"0.4.14","## 变更内容\n* 由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1012 中增强了追踪管理的相关文档和功能。\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1022 中解绑并升级了版本依赖。\n\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.13...0.4.14","2025-05-30T20:46:28",{"id":170,"version":171,"summary_zh":172,"released_at":173},71786,"0.4.13","## 变更内容\n* 由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F985 中回滚路线图\n* 由 @devin-ai-integration 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F993 中更新 openai-agents SDK 笔记本，使其在安装依赖时静默进行，并添加重启内核的说明\n* 由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F983 中添加用于处理流式响应的流\n* 由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F987 中实现 LLM 库的懒加载，以防止重复监控\n* 由 @the-praxs 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F997 中回退“修复：LLM 调用重复 (#981)”\n* 由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1002 中在请求头中添加 user-agent\n* 由 @the-praxs 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1010 中对 `AG2` 进行小幅重构，调整 `LIBRARY_NAME` 和 `LIBRARY_VERSION`\n* 由 @the-praxs 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1005 中修复 Agents SDK 的工具调用、提示词响应及其他数据\n* 由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1008 中增强工具装饰器功能\n* 由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F992 中修复 CrewAI 根跨度中的标签问题\n* 由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1009 中为 AgentOps 实现 Google ADK 监控\n* 由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F999 中改进根跨度管理\n* 由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F1011 中将 pyproject.toml 中的版本号更新至 0.4.13\n\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.12...0.4.13","2025-05-27T22:32:15",{"id":175,"version":176,"summary_zh":177,"released_at":178},71787,"0.4.12","## 变更内容\n* 重构日志记录，使用内存缓冲区代替文件。由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F961 中完成。\n* 更新了文档中的示例。由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F973 中完成。\n* 添加了绿色勾选标记。由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F972 中完成。\n* 对象上传的响应类型无需再是 Pydantic 模型。由 @tcdent 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F971 中完成。\n* 笔记本 1。由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F978 中完成。\n* Agents SDK 示例。由 @areibman 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F979 中完成。\n* 为 Linitng 重构笔记本示例。由 @Dwij1704 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F980 中完成。\n* 修复 Agent 装饰器。由 @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F976 中完成。\n* 修复：重复的 LLM 调用。由 @the-praxs 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F981 中完成。\n\n## 新贡献者\n* @fenilfaldu 在 https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F976 中完成了首次贡献。\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.11...0.4.12","2025-05-15T19:59:01",{"id":180,"version":181,"summary_zh":182,"released_at":183},71788,"0.4.11","## What's Changed\r\n* Add httpx dependency to fix import error by @devin-ai-integration in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F966\r\n* fix versions by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F970\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.10...0.4.11","2025-05-12T20:38:04",{"id":185,"version":186,"summary_zh":187,"released_at":188},71789,"0.4.10","## What's Changed\r\n* Update README and example notebooks to reflect IBM Watsonx branding a… by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F949\r\n* Update naming by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F950\r\n* IO.net IO Intelligence integration document by @kernelzeroday in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F954\r\n* Enhance the wrap_llm_call function to set prompt and completion attri… by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F958\r\n* upgrade version by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F963\r\n\r\n## New Contributors\r\n* @kernelzeroday made their first contribution in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F954\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.9...0.4.10","2025-05-08T20:36:59",{"id":190,"version":191,"summary_zh":192,"released_at":193},71790,"0.4.9","## What's Changed\r\n* Introducing V2 Documentation with Comprehensive Updates by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F938\r\n* Logo fixes by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F942\r\n* Logo fixes by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F943\r\n* Logo fixes by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F944\r\n* Logo fixes by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F946\r\n* Add AG2Instrumentor by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F939\r\n* Add IBM watsonx.ai instrumentation by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F941\r\n* Bump version to 0.4.9 in pyproject.toml by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F948\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.8...0.4.9","2025-05-02T23:51:22",{"id":195,"version":196,"summary_zh":197,"released_at":198},71791,"0.4.8","## What's Changed\r\n* Removed import break by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F937\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.7...0.4.8","2025-04-27T09:10:07",{"id":200,"version":201,"summary_zh":202,"released_at":203},71792,"0.4.7","## What's Changed\r\n* Add system stats and imported libraries tracking by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F914\r\n* Add session management during shutdown and improve session handling by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F915\r\n* bad api key doesnt block by @bboynton97 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F919\r\n* Update Anthropic instrumentation by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F918\r\n* Update import LangchainCallbackHandler in documentation by @pavalucas in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F927\r\n* S3 logs by @bboynton97 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F928\r\n* Refactor CrewAI instrumentation to enhance span attribute management by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F926\r\n* Add GoogleGenerativeAIInstrumentor by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F932\r\n* Bump version to 0.4.7 by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F933\r\n\r\n## New Contributors\r\n* @pavalucas made their first contribution in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F927\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.6...0.4.7","2025-04-24T00:39:11",{"id":205,"version":206,"summary_zh":207,"released_at":208},71793,"0.4.6","## What's Changed\r\n* Fix token counts in OpenAI Agents. by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F888\r\n* Add LLM_CONTENT_COMPLETION_CHUNK attribute to SpanAttributes class by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F885\r\n* collapse details by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F893\r\n* Fix token counts in OpenAI Agents redux by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F890\r\n* Fix incorrect url in README.md by @mlkorra in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F896\r\n* Ensure auto started sessions behave properly by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F899\r\n* OpenAI Agents voice support by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F900\r\n* OpenAI Responses by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F892\r\n\r\n## New Contributors\r\n* @mlkorra made their first contribution in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F896\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.5...0.4.6\r\n\r\n## What's Changed\r\n* Fix token counts in OpenAI Agents. by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F888\r\n* Add LLM_CONTENT_COMPLETION_CHUNK attribute to SpanAttributes class by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F885\r\n* collapse details by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F893\r\n* Fix token counts in OpenAI Agents redux by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F890\r\n* Fix incorrect url in README.md by @mlkorra in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F896\r\n* Ensure auto started sessions behave properly by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F899\r\n* OpenAI Agents voice support by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F900\r\n* OpenAI Responses by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F892\r\n* bump to `0.4.6` by @the-praxs in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F911\r\n\r\n## New Contributors\r\n* @mlkorra made their first contribution in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F896\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.5...0.4.6","2025-04-07T22:18:26",{"id":210,"version":211,"summary_zh":212,"released_at":213},71794,"0.4.5","## What's Changed\r\n* Updated README with updated usage examples for new decorators and session management. Introduce `session`, `agent`, `operation`, `task`, and `workflow` decorators for improved observability. by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F862\r\n* Updated unit tests for decorators by adding workflow and task nesting validation. by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F863\r\n* Update README.md by @srilaasya in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F872\r\n* Update README.md by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F874\r\n* Configure Ruff to enforce unused imports and undefined names by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F881\r\n* Log deeplink to trace on AgentOps dashboard. by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F879\r\n* Add langchanin callback handler by @Dwij1704 in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F876\r\n* Version bump to 0.4.5 by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F880\r\n* Refactor Agents SDK instrumentation.  by @tcdent in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F854\r\n\r\n## New Contributors\r\n* @srilaasya made their first contribution in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F872\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.4...0.4.5","2025-03-25T00:05:29",{"id":215,"version":216,"summary_zh":217,"released_at":218},71795,"0.4.4","# Changelog\r\n- [x] Reduce the delay of span export from five seconds to one second to avoid orphaned spans during unexpected system exits (not necessarily exceptions but explicit exits from other applications). Expose this configuration parameter as an environment variable\r\n- [x] Fixes #838 @the-praxs \r\n- [x] Fixes #710 @tcdent \r\n- [x] Restores support for Python 3.9 which has not yet reached [EOL](https:\u002F\u002Fdevguide.python.org\u002Fversions\u002F)\r\n- [x] Crew released version 105 which interacts with the legacy public API in a [different manner](https:\u002F\u002Fgithub.com\u002FcrewAIInc\u002FcrewAI\u002Fpull\u002F2048). AgentOps is now integrated as a separate [module] (https:\u002F\u002Fgithub.com\u002FcrewAIInc\u002FcrewAI\u002Fblob\u002Fmain\u002Fsrc\u002Fcrewai\u002Futilities\u002Fevents\u002Fthird_party\u002Fagentops_listener.py) and we have included support for this refactor here.\r\n- [x] Versions less than 105 were mostly covered by the existing implementation is also supported in this PR.\r\n- [x] Tested with `openai-agents` to verify span export \r\n\r\n# Tests\r\n- [x] Implements [tests\u002Funit\u002Ftest_session_legacy.py](https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fblob\u002F0.4.4\u002Ftests\u002Funit\u002Ftest_session_legacy.py) to verify full backwards compatibility with CrewAI's and other libs\r\n- [x] Removes deprecated failing tests:\r\n  - `test_insturmentation.py`\r\n  - `test_instrumentation_errors.py`\r\n  - `test_core.py`\r\n- [x] Favors `test_decorators.py`\r\n- [x] Adds an implementation of crew AI testing in `test\u002Funit\u002Ftest_session_legacy.py`\r\n","2025-03-17T21:07:47",{"id":220,"version":221,"summary_zh":222,"released_at":223},71796,"0.4.3","## What's Changed\r\n* fix-a-lot by @teocns in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F830\r\n* replace icons by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F831\r\n* Better icons by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F832\r\n* Update pyproject.toml by @areibman in https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fpull\u002F836\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FAgentOps-AI\u002Fagentops\u002Fcompare\u002F0.4.2...0.4.3","2025-03-14T17:35:26",{"id":225,"version":226,"summary_zh":227,"released_at":228},71797,"0.4.2","fixes compatibility with crewAI.","2025-03-13T16:55:58",[230,239,247,255,263,275],{"id":231,"name":232,"github_repo":233,"description_zh":234,"stars":235,"difficulty_score":236,"last_commit_at":237,"category_tags":238,"status":95},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",[78,77,79],{"id":240,"name":241,"github_repo":242,"description_zh":243,"stars":244,"difficulty_score":65,"last_commit_at":245,"category_tags":246,"status":95},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,"2026-04-05T11:33:21",[78,79,80],{"id":248,"name":249,"github_repo":250,"description_zh":251,"stars":252,"difficulty_score":65,"last_commit_at":253,"category_tags":254,"status":95},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",[78,77,79],{"id":256,"name":257,"github_repo":258,"description_zh":259,"stars":260,"difficulty_score":65,"last_commit_at":261,"category_tags":262,"status":95},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",[78,80],{"id":264,"name":265,"github_repo":266,"description_zh":267,"stars":268,"difficulty_score":65,"last_commit_at":269,"category_tags":270,"status":95},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",[77,271,272,273,79,76,80,78,274],"数据工具","视频","插件","音频",{"id":276,"name":277,"github_repo":278,"description_zh":279,"stars":280,"difficulty_score":236,"last_commit_at":281,"category_tags":282,"status":95},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",[79,77,78,80,76]]