Composio logo

Composio

Unified SDKs to empower AI agents with real‑world tools

Python and TypeScript SDKs that let agents call 500+ app APIs through Composio, with built‑in support for major AI frameworks.

Composio banner

Overview

Overview

The Composio SDK provides official Python and TypeScript libraries that enable developers to equip autonomous agents with live data from over 500 SaaS applications. By exposing a consistent, async‑friendly API, the SDK lets agents retrieve tools, invoke actions, and return structured results without writing custom integration code.

Capabilities

Both the TypeScript (@composio/core) and Python (composio) packages include core client functionality and a suite of provider extensions for OpenAI, Anthropic, LangChain, Gemini, and many others. Developers can install language‑specific packages via npm or pip, import the client, and fetch toolkits (e.g., Hackernews) to power agents built with OpenAI Agents, LangChain, CrewAI, or AutoGen. The SDK is type‑safe in TypeScript and follows modern Pythonic conventions, supporting Python 3.10+ and async/await patterns.

Deployment

Installation is straightforward (npm install @composio/core or pip install composio). The libraries run in Node.js, browser, and serverless environments, and can be extended with custom providers to integrate any AI framework. For broader MCP use, the companion Rube server connects AI tools to the same ecosystem of apps.

Highlights

Type‑safe TypeScript client for Node.js and browser environments
Pythonic async API supporting Python 3.10+
Provider packages for OpenAI, Anthropic, LangChain, Gemini, and more
Unified tool retrieval enabling access to 500+ third‑party services

Pros

  • Consistent API across languages reduces learning curve
  • Extensive provider ecosystem covers most popular LLM frameworks
  • Async design fits modern agent runtimes
  • OpenAPI‑driven generation keeps SDKs in sync with backend

Considerations

  • Vercel AI SDK support is missing in Python
  • Some providers (e.g., Mastra, Cloudflare) unavailable in Python
  • External API key management adds credential overhead
  • Documentation spans multiple package repos, requiring navigation

Managed products teams compare with

When teams consider Composio, these hosted platforms usually appear on the same shortlist.

CrewAI logo

CrewAI

Multi-agent automation framework & studio to build and run AI crews

LangGraph logo

LangGraph

Open-source framework for building stateful, long-running AI agents

Relevance AI logo

Relevance AI

No-code platform to build a team of AI agents with rich integrations

Looking for a hosted option? These are the services engineering teams benchmark against before choosing open source.

Fit guide

Great for

  • Developers building autonomous agents that need live data
  • Teams standardizing on a single SDK for both JavaScript and Python projects
  • Projects requiring integration with many SaaS apps via Composio
  • Experimentation with multiple LLM providers without changing code

Not ideal when

  • Purely offline AI workloads with no external API calls
  • Environments where adding npm or pip dependencies is prohibited
  • Use cases needing Vercel AI SDK support in Python
  • Applications requiring built‑in UI components rather than backend SDKs

How teams use it

News summarization agent

Agent fetches latest Hackernews posts via Composio tools and returns concise summaries.

CRM automation bot

Bot creates contacts in Salesforce and logs follow‑up tasks in Notion using Composio’s 500+ integrations.

Email triage assistant

Agent reads incoming Gmail messages and drafts replies, leveraging Composio’s Gmail tool.

Project management helper

Agent adds tickets to Jira and updates Asana tasks, orchestrated through the SDK.

Tech snapshot

TypeScript63%
Python23%
MDX12%
Shell1%
CSS1%
JavaScript1%

Tags

aiaiagentsllmfunction-callingagentic-aiagentsmcppythonremote-mcp-serverjsai-agentsdeveloper-toolstypescriptjavascriptgpt-4ssellmops

Frequently asked questions

Which programming languages are supported?

The official SDKs are available for TypeScript (Node.js and browsers) and Python 3.10+.

How do I add support for a new AI provider?

You can create a custom provider package following the SDK’s provider interface; documentation provides a guide.

Do I need an API key for every integrated app?

Composio requires a single API key for its platform; individual app authorizations are handled through the Composio UI.

Is the SDK compatible with async frameworks?

Both the TypeScript and Python SDKs expose async methods, making them suitable for async runtimes.

Can I use the SDK in serverless environments?

Yes, the SDKs are lightweight and can be imported in serverless functions such as AWS Lambda or Vercel Edge Functions.

Project at a glance

Active
Stars
26,411
Watchers
26,411
Forks
4,408
LicenseMIT
Repo age1 year old
Last commit8 hours ago
Primary languageTypeScript

Last synced 1 hour ago