Swagger MCP

Swagger MCP

Automatically converts Swagger/OpenAPI specifications into MCP servers, enabling AI agents to interact with any REST API through natural language by exposing endpoints as AI-friendly tools.

Category
Visit Server

README

Swagger MCP

Automatically convert a Swagger/OpenAPI specification into an MCP server for use with Windsurf, Cursor, or other tools.

Quickstart

swagger-mcp uses pipx

Install from PyPI using pipx (recommended):

brew install pipx
pipx ensurepath
pipx install --force swagger-mcp

Alternatively, install from source:

git clone https://github.com/context-labs/swagger-mcp.git
cd swagger-mcp
pipx install -e . --force

Confirm the installation succeeded:

which swagger-mcp
which swagger-mcp-sample-server

Spin up a sample "products and product categories" API on your local machine on port 9000:

swagger-mcp-sample-server

Visit http://localhost:9000/docs to confirm the sample server is running.

We'll use this sample server to show how to configure an MCP server in Windsurf.

Make sure the sample server is running before following the Windsurf or Cursor instructions below.

Windsurf

Start an MCP Server in Windsurf (Windsurf Settings -> Settings -> Cascade -> Manage Plugins -> View Raw Config):

{
  "mcpServers": {
    "product-mcp": {
      "command": "swagger-mcp",
      "args": [
        "--spec",
        "http://localhost:9000/openapi.json",
        "--name",
        "Product-MCP",
        "--server-url",
        "http://localhost:9000"
      ]
    }
  }
}

That's it! Your API is now accessible through Windsurf, Cursor, or other tools as a set of AI-friendly tools.

Ask your AI agent to list, create, update, and delete products and categories.

Demo

Cursor (>=v0.46)

Support for Cursor is still in beta as Cursor MCP integration matures. Windsurf is currently the preferred experience. To add a custom MCP server in Cursor, go to Settings -> Cursor Settings -> Tools & Integrations -> Add Custom MCP.

{
  "mcpServers": {
    "product mcp": {
      "command": "swagger-mcp",
      "args": [
        "--spec",
        "http://localhost:9000/openapi.json",
        "--name",
        "Product-MCP",
        "--server-url",
        "http://localhost:9000",
        "--cursor"
      ]
    }
  }
}

Please Note: In Cursor, you may need to replace the command swagger-mcp with the full path to the swagger-mcp executable, which you can find by running which swagger-mcp.

Also note the --cursor flag. This is for Cursor compatibility.

Again, MCP integration is currently in beta in Cursor as of v0.46 and may not work as expected. Currently, Windsurf is a better experience in general.

See other examples in Other Fun Servers.

Additional Options

  1. You can pass a JSON file, YAML file, or URL for the --spec option:

    • /path/to/openapi.json
    • /path/to/openapi.yaml
    • https://api.example.com/openapi.json
  2. Filter endpoints: Only include endpoints where the path matches the regex pattern:

{
  "mcpServers": {
    "product mcp": {
      "command": "swagger-mcp",
      "args": [
        "--spec",
        "http://localhost:9000/openapi.json",
        "--name",
        "product-mcp",
        "--server-url",
        "http://localhost:9000",
        "--include-pattern",
        "category"
      ]
    }
  }
}
  1. Filter endpoints: Exclude endpoints where the path matches the regex pattern:
{
  "mcpServers": {
    "product mcp": {
      "command": "swagger-mcp",
      "args": [
        "--spec",
        "http://localhost:9000/openapi.json",
        "--name",
        "product-mcp",
        "--server-url",
        "http://localhost:9000",
        "--exclude-pattern",
        "product"
      ]
    }
  }
}
  1. Authentication
{
  "mcpServers": {
    "product mcp": {
      "command": "swagger-mcp",
      "args": [
        "--spec",
        "http://localhost:9000/openapi.json",
        "--name",
        "product-mcp",
        "--server-url",
        "http://localhost:9000",
        "--bearer-token",
        "your-token-here",
      ]
    }
  }
}
  1. Custom headers
{
  "mcpServers": {
    "product mcp": {
      "command": "swagger-mcp",
      "args": [
        "--spec",
        "http://localhost:9000/openapi.json",
        "--name",
        "product-mcp",
        "--server-url",
        "http://localhost:9000",
        "--header",
        "X-Some-Header:your-value",
        "--header",
        "X-Some-Other-Header:your-value",
      ]
    }
  }
}
  1. Server URLs If the OpenAPI spec already contains a specific server URL, you don't have to provide it as a command line argument. But if you do, the command line --server-url overrides all endpoints.

  2. Constant Values

If you want to always automatically provide a value for a parameter, you can use the --const option. You can include as many --const options as you need.

{
  "mcpServers": {
    "product mcp": {
      "command": "swagger-mcp",
      "args": [
        "--spec",
        "http://localhost:9000/openapi.json",
        "--name",
        "product-mcp",
        "--server-url",
        "http://localhost:9000",
        "--const",
        "parameter-name:your-value",
        "--const",
        "parameter-name2:your-value2"
      ]
    }
  }
}

Supported Features

  • All HTTP methods (GET, POST, PUT, DELETE, etc.)
  • Path parameters
  • Query parameters
  • Textual Multi-Part Request Body Fields
  • JSON Request body
  • Bearer Token Authentication
  • Custom Headers
  • Constant Values

Limitations

  • Endpoints that have recursive schema references are not yet supported.
  • Cursor MCP integration is very early and frankly broken. I try to address some of these with cursor mode --cursor, but it's still not great. Until Cursor MCP support gets better, you'll be happier with Windsurf.
  • We will never support automatic OAuth workflow execution. If the OAuth workflow creates a bearer token, you must obtain this token yourself by performing OAuth out-of-band, and provide this bearer token as a command line argument.
  • We do not support Swagger/OpenAPI specifications spread across multiple files (i.e.; fragments, extensions, etc.).
  • We do not support path variable substitution in server URLs (but we do support path variables in endpoints).
  • In general, we do not support all Swagger/OpenAPI features. The Swagger/OpenAPI standard is vast, and support for more obscure features will be added as needed / requested.

Help

  • If you have trouble spinning up a server, try the following command: REAL_LOGGER=true swagger-mcp-parse-dry-run ... and provide all the same arguments you would use to spin up a server. Include this information in any issue you file.
  • If you find a Swagger API specification that is not supported and you can't use any of the available parameters for a workaround, please file an issue. We will add support for it as needed / requested.

Command Line Options

  • --spec (required): Path or URL to your OpenAPI/Swagger specification
  • --name (required): Name for your MCP server (shows up in Windsurf/Cursor)
  • --server-url: Base URL for API calls (overrides servers defined in spec)
  • --bearer-token: Bearer token for authenticated requests
  • --additional-headers: JSON string of additional headers to include in all requests
  • --include-pattern: Regex pattern to include only specific endpoint paths (e.g., "/api/v1/.*")
  • --exclude-pattern: Regex pattern to exclude specific endpoint paths (e.g., "/internal/.*")
  • --header: key:value pair of an extra header to include with all requests. Can be included multiple times to specify multiple headers.
  • --const: key:value pair of a constant value to always use for a parameter, if the parameter is present on the endpoint (can be a path variable, query parameter, top-level request body property, or multi-part non-file form data field). Can be included multiple times to specify multiple const values.
  • --cursor: Run in cursor mode

Authentication

For APIs requiring authentication:

{
  "mcpServers": {
    "product mcp": {
      "command": "swagger-mcp",
      "args": [
        "--spec",
        "http://localhost:9000/openapi.json",
        "--name",
        "product-mcp",
        "--server-url",
        "http://localhost:9000",
        "--bearer-token",
        "your-token-here",
        "--cursor"
      ]
    }
  }
}
{
  "mcpServers": {
    "product mcp": {
      "command": "swagger-mcp",
      "args": [
        "--spec",
        "http://localhost:9000/openapi.json",
        "--name",
        "product-mcp",
        "--server-url",
        "http://localhost:9000",
        "--header",
        "X-API-Key:your-key",
        "--cursor"
      ]
    }
  }
}

Other Fun Servers

Countries

{
  "mcpServers": {
    "countries": {
      "command": "swagger-mcp",
      "args": [
        "--spec",
        "https://restcountries.com/openapi/rest-countries-3.1.yml",
        "--name",
        "countries",
        "--server-url",
        "https://restcountries.com/",
        "--const",
        "fields:name"
      ]
    }
  }
}

Petstore

{
  "mcpServers": {
    "petstore": {
      "command": "swagger-mcp",
      "args": [
        "--spec",
        "https://petstore.swagger.io/v2/swagger.json",
        "--name",
        "petstore",
        "--server-url",
        "https://petstore.swagger.io/v2",
        "--header",
        "X-API-Key:special-key",
      ]
    }
  }
}

For Developers

Installation

For development, install with development dependencies:

# Clone the repository
git clone https://github.com/context-labs/swagger-mcp.git
cd swagger-mcp

# Install in development mode with dev dependencies
pip install -e ".[dev]"

For a global installation on your local machine, use:

bash scripts/install-global.sh

Unit Tests

pytest swagger_mcp/tests/unit -v

Integration Tests

pytest swagger_mcp/tests/integration -v --capture=no --log-cli-level=INFO

MCP Inspector (For interactive exploration of the MCP Server)

You'll have to do a global installation of your latest code first (bash scripts/install-global.sh), then you can run the inspector script.

You'll see the server type STDIO and the command swagger-mcp pre-filled.

bash scripts/inspector.sh

Click "Connect" and then "List Tools" to begin interacting with your MCP Server.

MCP Inspector

Logging

To run the server with logs enabled, set the REAL_LOGGER environment variable to true:

REAL_LOGGER=true swagger-mcp --spec http://localhost:9000/openapi.json --name product-mcp --server-url http://localhost:9000

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured