From API to MCP server
in 6 seconds.
Describe any API. Get a production-ready MCP server
with smart caching, retry, rate limiting — instantly.
What you get
Every generated MCP server ships with these out of the box.
Smart Runtime
Cache, retry with exponential backoff, rate limit detection, auto-pagination. Every generated server ships with production-grade resilience built in.
Auto-Discovery
Point at any OpenAPI spec URL and Forgem discovers all endpoints automatically. No manual endpoint listing needed.
Production-Ready
TypeScript, proper error handling, auth support (Bearer, Basic, API-Key). Built, compiled, and ready to ship.
Zero Config
Auto-configures Claude Code and Cursor. One command generates everything. No JSON editing, no documentation hunting.
Three steps. Six seconds.
From any API to a working MCP server your AI can use.
Describe your API
Give Forgem a name, base URL, and endpoints — or just point it at an OpenAPI spec URL. That's all it needs.
Forgem generates your MCP server
In under 6 seconds, Forgem produces a complete MCP server with smart runtime: caching, retry with backoff, rate limit detection, and auto-pagination.
Paste the config. Done.
Forgem outputs a settings.json snippet. Paste it into Claude Code or Cursor. Your AI assistant can now call every endpoint of your API.
Pricing
Start free. Scale when you need more.
Free
- 15 generations/month
- Smart runtime included
- OpenAPI auto-discovery
- Community support
Pro
- Unlimited generations
- Priority generation queue
- Private MCP servers
- Advanced auth patterns
- Priority support
Start generating.
Describe your API. Get a production-ready MCP server. Six seconds.