Getting Started

Step 1: install the CLI

Follow Installation first.

After installation, confirm the command is available:

openapi-to-mcp --help

Step 2: choose the command you need

Use generate when you want a reusable generated project on disk.

openapi-to-mcp generate \
  --openapi-json ./openapi.yaml \
  --output-dir ./generated-server

Use run when you want one command to generate, build, and start a server locally.

openapi-to-mcp run \
  --openapi-json https://petstore.swagger.io/v2/swagger.json \
  --target-api-base-url https://petstore.swagger.io/v2

Use test-server when you want to smoke-test a running server.

openapi-to-mcp test-server \
  --transport streamable-http \
  --host 127.0.0.1 \
  --port 8080 \
  --mcp-endpoint /mcp \
  --list-tools

Step 3: know when Node.js is required

• generate: does not need Node.js
• run: needs node and npm
• generated projects themselves need Node.js to build and run

Step 4: go deeper

• generate
• run
• test-server
• Auth and Environment
• Examples

Development and source workflows

Source checkout, uv sync, just, hooks, and helper scripts are intentionally documented separately in Local Workflows.