Composio is a platform that provides Software Development Kits for Python and TypeScript to enable AI agents to interact with external tools and services. The repository contains the official SDKs that integrate with various agentic frameworks and libraries, allowing developers to build AI agents capable of turning user intent into concrete actions across 1000+ toolkits. The platform handles tool search, context management, authentication, and provides a sandboxed workbench environment for agent execution.

The TypeScript SDK is designed for both Node.js and browser environments with full type-safe support and comprehensive type definitions. The Python SDK offers a Pythonic interface supporting Python 3.10 and above, following modern Python development practices. Both SDKs pull OpenAPI specifications from the Composio backend to generate up-to-date API documentation automatically during the build process. The repository is written primarily in TypeScript and maintains extensive provider support across major AI frameworks including OpenAI, Anthropic, LangChain, LangGraph, LlamaIndex, Google Gemini, Vercel AI SDK, Mastra, and Cloudflare Workers AI for TypeScript, with additional support for CrewAI and AutoGen in Python.

The project is organized into multiple npm and PyPI packages. Core packages include @composio/core for TypeScript and composio for Python. Provider-specific packages exist for each supported framework, such as @composio/openai, @composio/langchain, and composio-crewai. Utility packages like @composio/json-schema-to-zod and @composio/ts-builders provide additional development tools. The repository also includes Rube, a Model Context Protocol server built with Composio that connects AI tools to over 500 apps including Gmail, Slack, GitHub, and Notion, integrating with AI clients like Cursor and Claude Desktop.

According to GitGenius tracking data, the repository has grown modestly with stargazers increasing from 29,089 to 29,090 between checks. The project maintains active issue and pull request management with a median response latency of 0.9 hours across 616 tracked items, though the mean latency is 105.7 hours. The most frequently applied issue labels are bug with 112 occurrences, stale with 39, and enhancement with 27. The most active contributors tracked by GitGenius are Uday-sidagana with 242 events, haxzie with 218 events, and abhishekpatil4 with 206 events. The repository shares overlapping contributors with prisma/prisma, nousresearch/hermes-agent, and letta-ai/letta, indicating cross-project collaboration within the AI agent development ecosystem.

The platform is classified across multiple domains including LLM Integration, Tool Orchestration, Workflow Automation, Application Connectors, AI Agents, Unified API, Extensible Framework, Actionable AI, Software Integration, and Automation Platform. The project welcomes community contributions and provides contribution guidelines for developers interested in extending the SDKs. Documentation is available at docs.composio.dev, and the project is licensed under the MIT License.