server
MCPfinder is a Node.js tool designed to help AI agents expand their capabilities by interacting with MCP (Model Context Protocol) servers. It allows language models to discover and manage new capabilities autonomously, aiding in transforming them into more powerful, evolving agents.
MCPfinder 🔧🤖
App Store for Agents
MCPfinder is a tool for enabling language models to install new capabilities via the MCP protocol, acting as a map and toolbox to transform static models into evolving agents. It supports applications such as Cursor, Claude Desktop, and Windsurf, allowing AI assistants to discover and manage MCP servers via functionalities like search_mcp_servers and add_mcp_server_config. It provides a user-friendly approach to enhance AI capabilities without manual coding.
Features:
- Framework for AI agents to autonomously expand their own capabilities by discovering MCP servers.
- Support for running locally alongside applications supporting the MCP protocol.
- Tools for searching and managing MCP servers via a registry.
Usage:
- Run an interactive setup tool with
npx -y @mcpfinder/server installto configure the MCP setup for your client application. - The application can be run in Stdio or HTTP mode, and supports command executions like
installandregisterfor initial setup or server registration.
Tools Provided:
search_mcp_servers: To search the MCPfinder Registry for available servers.get_mcp_server_details: To retrieve detailed information about specific servers.add_mcp_server_config: To add or update configuration for a specific MCP server.remove_mcp_server_config: To remove server configurations.
Security Considerations
The server modifies local configuration files, relying on client-side security enforcement.