MCP Client (Node.js)

Example Use Cases

• 1

  Connect to any MCP-compliant server and automatically discover and invoke server-provided tools.

• 2

  Route LLM calls through an OpenAI-compatible model provider with function-calling support and handle tool results in a structured loop.

• 3

  Comprehensive JSON-formatted logging of LLM requests, tool executions, and errors for auditing and debugging.

Description

This project provides a modular MCP client for Node.js that connects to any MCP-compliant server, discovers available tools automatically, and mediates LLM-driven tool calls. It is compatible with OpenAI-style APIs, includes an interactive CLI, and logs all LLM requests, tool invocations, and errors to disk in JSON format. The client simplifies building agents that need structured tool execution, integrates with existing MCP servers, and provides robust auditing and debugging capabilities.