MigrateForce Documentation

Convert any REST API into an MCP server in under five minutes. Run locally, enrich via cloud, and ship agent-ready tooling.

npx @migrateforce/cli convert \
  --spec openapi.yaml \
  --base-url https://api.example.com \
  --output ./my-mcp-server

That command reads your OpenAPI spec, maps every endpoint to an MCP tool, and writes a deployable Python server to disk. No account required.

Quick Start

Install the CLI, point it at your OpenAPI spec, and get a working MCP server in one command. Five-minute walkthrough.

Start here

Local vs Cloud

Understand what runs on your machine and what the cloud adds: linting, readiness scores, skill matching, and project tracking.

Learn more

Skills

Reusable knowledge files that tell AI agents what to do with your MCP tools. Browse the catalog, install from GitHub, or author your own.

Explore skills

Web Dashboard

Upload specs, edit tool mappings visually, run the Migration Agent, and download generated code — all from the browser.

View guide

What is MigrateForce?

MigrateForce converts REST APIs into Model Context Protocol (MCP) servers that AI agents — Claude, GPT, Gemini — can call as tools. You give it an OpenAPI spec and a base URL. It gives you a deployable Python server with typed tool definitions, auth handling, and a Docker setup.

How It Works

1

Parse

Your OpenAPI 3.0 or Swagger 2.0 spec is parsed and validated. Every endpoint is extracted with its method, path, parameters, and request body.

2

Map

Each endpoint gets a semantic tool name (get_user_profile, create_order) and a description written for LLM consumption — precise enough that an agent can decide when and how to call it.

3

Generate

A complete MCP server is generated: main.py, Dockerfile, requirements.txt, README.md, .env.example, and mcp-manifest.json. Ready to run.

4

Enrich (optional, cloud)

Add --cloud to get readiness linting, improvement advice, and skill catalog matching from the MigrateForce API. No generated code leaves your machine — enrichments are analysis only.

Two ways to use MigrateForce
CLI — fastest path. One command, files on disk, done. Start here.
Web Dashboard — visual editor for tool mappings, agent-powered analysis, project history. Open dashboard guide.

What Ships Today

CLI (@migrateforce) — convert (OpenAPI → MCP server), lint (readiness check), skills (catalog browsing and install). Supports --cloud for API-backed enrichments and --json for CI pipelines.

Skills CLI (@migrateforce/skills) — find, install, validate, and compose SKILL.md files that give AI agents structured knowledge about your domain.

Web Dashboard — upload OpenAPI specs, review and edit tool mappings, run the Migration Agent (Claude-powered spec analysis), generate and download MCP server code.

Cloud API — authenticated endpoints for linting, skill matching, and project management. The CLI's --cloud flag calls these automatically. Code generation always happens locally.