Skip to content

Getting Started

This guide gets you from zero to a running MCP server in under five minutes.

Prerequisites

  • Python 3.11 or later
  • An OpenAPI spec file (or any supported format — see supported formats)

1. Install

pip install api2mcp

Verify the installation:

api2mcp --version

2. Generate Your MCP Server

Point api2mcp generate at any OpenAPI spec:

api2mcp generate openapi.yaml --output ./my-server

This produces a ./my-server/ directory containing:

my-server/
├── spec.yaml          # Validated copy of your spec
├── tools.py           # Generated MCP tool definitions
└── server.py          # Runnable MCP server entry point

Using the interactive wizard

If you prefer guided setup, run api2mcp wizard instead for a step-by-step interactive experience.

3. Start the Server

api2mcp serve ./my-server

The MCP server starts on http://127.0.0.1:8000 by default (Streamable HTTP transport).

INFO     Starting MCP server on http://127.0.0.1:8000
INFO     Transport: http (Streamable HTTP)
INFO     Tools loaded: 12

4. Connect an LLM Client

Point your MCP client at http://127.0.0.1:8000. All endpoints from your OpenAPI spec are now available as named MCP tools — no further configuration needed.

5. (Optional) Validate First

Before generating, you can validate your spec to catch errors early:

api2mcp validate openapi.yaml

Use --strict to treat warnings as errors:

api2mcp validate openapi.yaml --strict

Supported Input Formats

Format Command Notes
OpenAPI 3.0 / 3.1 api2mcp generate spec.yaml Most common
Swagger 2.0 api2mcp generate swagger.yaml Auto-migrated to 3.0
GraphQL SDL api2mcp generate schema.graphql Requires pip install "api2mcp[graphql]"
Postman Collection api2mcp generate collection.json v2.0 and v2.1

Configuration File

Create .api2mcp.yaml in your project root to set defaults:

output: ./generated
host: 127.0.0.1
port: 8000
transport: http    # or stdio
log_level: info

CLI flags always override the config file.

Next Steps