SDK Overview
SDK Packages
TypeScript SDK
@ag-kit/agents
Built-in agents and adapters for LangGraph, LangChain, OpenAI, and CrewAI
@ag-kit/server
Node.js server for AI agents
@ag-kit/tools
Unified tool integration for code execution, file operations, and more
@ag-kit/shared
Shared TypeScript types, utilities, and common interfaces
Python SDK
ag_kit.agents
Python implementation of agent adapters and frameworks
ag_kit.server
Python server for hosting and managing AI agents
ag_kit.tools
Python tool integrations and utilities
Headless UI Components
@ag-kit/ui-react
React components for building agent interfaces
@ag-kit/ui-miniprogram
WeChat Mini Program components for mobile agent interfaces
Quick Start
TypeScript Installation
Python Installation
UI Components Installation
Key Features
- Multi-Framework Support - Works with LangGraph, LangChain, OpenAI, and CrewAI
- Cross-Platform - TypeScript and Python implementations
- Headless UI - Framework-agnostic UI components
- Tool Integration - Unified interface for various tools and services
- Type Safety - Full TypeScript support with comprehensive type definitions
- Scalable Architecture - Built for production-ready agent systems
- Real-time Communication - WebSocket support for interactive agents
- Extensible - Plugin architecture for custom agents and tools
Getting Started
- Choose your framework - Select from LangGraph, LangChain, OpenAI, or CrewAI
- Install the SDK - Use TypeScript or Python based on your preference
- Create an agent - Use built-in adapters or create custom agents
- Add tools - Integrate code execution, file operations, and more
- Build the UI - Use React or Mini Program components
- Deploy - Use the server package to host your agents
For detailed guides and examples, check out our Getting Started section or explore the Playground for interactive examples.