Smithery is a registry and management platform for Model Context Protocol (MCP) servers, giving developers a central hub to discover, install, deploy, and manage a large and fast-changing catalog of integrations for AI agents. It solves the challenge of finding and deploying MCP servers by combining a searchable registry, CLI-based setup, hosted deployment options, and a Platform API for discovery, deployments, connections, namespaces, and scoped access tokens. The durable value is not an exact catalog count but the operational layer around MCP server discovery and connection management.
Smithery supports both local installations where tools run on the user machine and hosted solutions where tools run on Smithery infrastructure, giving developers flexibility in how they deploy and manage their MCP servers. The platform provides a CLI tool for installing and managing servers, detailed documentation for each listed server, usage analytics, and compatibility information for different AI clients. Smithery also offers an API for programmatic access to the registry, enabling AI platforms and agent frameworks to integrate MCP server discovery and deployment directly into their workflows.
Smithery targets AI developers, agent builders, and platform teams who need to equip their AI applications with external tool capabilities through the Model Context Protocol. It integrates with major MCP clients including Claude, Cursor, Windsurf, and Cline, providing tailored installation instructions for each platform. The registry is particularly valuable for developers who want to quickly prototype AI agent capabilities by browsing available tools, testing them through Smithery hosted infrastructure, and then deploying proven configurations to their production environments with the Smithery CLI.
