LangGraph Multi-Agent Boilerplate
This LangGraph Multi-Agent Boilerplate provides a FastAPI-based framework for building AI agent clusters using a supervisor architecture, where crews of specialized agents collaborate to accomplish complex tasks through LangGraph workflow orchestration. Built with PostgreSQL for state management, OpenRouter API integration for multiple AI models, and MCP server support for tool calling via langchain-mcp-adapters, it enables agents to communicate with each other while the supervisor analyzes user inputs, creates detailed execution plans, delegates tasks to appropriate agents, and synthesizes results into coherent responses. The implementation features streaming chat endpoints, activity logging, Cloudflare R2 storage integration, and comprehensive CRUD APIs for managing crews, agents, conversations, and MCP tools, making it valuable for developers who want to rapidly prototype multi-agent AI systems without building the foundational infrastructure from scratch.
