APIFold

APIFold is an AI developer tool that converts OpenAPI specifications into Model Context Protocol (MCP) endpoints so AI agents can call existing APIs without custom server code. It transforms OpenAPI 3.0, 3.1, and Swagger 2.0 specs into callable MCP tools with typed parameters, then exposes them over an SSE plus JSON-RPC protocol for real-time communication. The generated tools work across AI coding agents including Claude Desktop, Cursor, GitHub Copilot, Windsurf, and Continue. A visual dashboard lets you manage specs, tools, connections, and authentication configurations in one place, while security features include AES-256 encryption, credential vaulting, CSRF protection, and request validation. APIFold can run cloud-hosted or self-hosted via Docker Compose, and its codebase is open source (AGPL-3.0 and MIT licensed) on GitHub with 80%+ test coverage. It is aimed at developers and enterprises that want to wire their REST APIs into agentic workflows quickly. By treating an OpenAPI spec as the single source of truth, APIFold removes the need to hand-write and maintain an MCP server for each integration, making it a practical bridge between traditional API documentation and the emerging MCP tool ecosystem.

Turning OpenAPI/Swagger specs into MCP tools, connecting REST APIs to AI coding agents (Cursor, Claude Desktop, GitHub Copilot, Windsurf, Continue), managing API specs and authentication from a dashboard, and self-hosting an API-to-MCP bridge via Docker.