# mcp-client

An example application pairing a LangGraph agent that connects to MCP servers and calls their tools with a CopilotKit frontend for UI and state synchronization.

This project is a two-part example: an agent folder holding a LangGraph agent that connects to MCP servers and calls their tools, and an app folder holding a frontend built with CopilotKit for UI and state synchronization with the agent.

It requires OpenAI and LangSmith API keys in .env files at the project root and inside the agent folder, and uses Poetry for the agent service's Python dependencies. The frontend and agent can be run in separate terminals for debugging, or together with a single command, serving the UI at localhost:3000.

Built with a TypeScript frontend using CopilotKit and a Python LangGraph agent managed with Poetry.

Website: https://mcp-client-five.vercel.app