MCPfinder

MCPfinder is a free, open-source discovery and installation layer designed for AI agents to locate and configure Model Context Protocol (MCP) servers. It serves as an MCP server itself, enabling AI assistants to programmatically search multiple MCP registries, inspect trust signals, and generate install-ready JSON configurations for downstream MCP servers. The tool is intended primarily for AI agents, with humans only needing to install it once; after installation, the AI assistant can use MCPfinder repeatedly to expand its capabilities on demand.

Key features include multi-registry search, aggregating MCP servers from the Official MCP Registry, Glama, and Smithery into a unified search interface. MCPfinder allows AI agents to find candidate servers by keyword, technology, or use case, inspect trust signals and required environment variables, and generate client-specific configuration snippets suitable for platforms such as Claude Desktop, Cursor, Claude Code, Cline, or Windsurf. The tool provides structured outputs, including confidence scores, recommendation reasons, warning flags, and install complexity, helping AI assistants make informed decisions about which servers to recommend or install.

MCPfinder is delivered as an MCP server that can be added to any MCP-compatible AI client. Installation is flexible, supporting direct execution via NPX (no install needed), global installation with NPM, or project-based integration. Once installed, it operates as an always-up-to-date agent-facing layer, bootstrapping from published snapshots and syncing from upstream registries. This ensures that AI agents have access to the latest available MCP servers and their metadata, with freshness signals provided by snapshot manifests.

0 and is free to use, with the source code available for inspection, contribution, or forking. MCPfinder is built by the community for the community, and as of its latest published snapshot, it aggregates tens of thousands of MCP servers across multiple registries. Its focus on agent-native discovery, trust signal inspection, and install-ready configuration distinguishes it as an AI-centric solution within the ecosystem of MCP tools.