Decorative pattern

Decorative pattern

ckanthony/openapi-mcp

Dockerized MCP server that enables AI agents to access any API with OpenAPI documentation through the MCP protocol.

Features

Dockerized MCP Server: Easily deployable using Docker, with a pre-built image available on Docker Hub or buildable locally.
OpenAPI/Swagger Integration: Reads OpenAPI (swagger.json or openapi.yaml) files and generates a Model Context Protocol (MCP) toolset for API interaction.
AI Agent Compatibility: Allows MCP-compatible clients (such as Cursor) to access any API described by OpenAPI, enabling AI agents to interact with APIs without extra coding.
Automatic Tool Definition Generation: Generates MCP tool definitions directly from OpenAPI specification files.
API Key Configuration: Supports passing API keys via command line flags, environment variables, or .env files.
Flexible API Key Handling: Configure API key parameter name and location (header, query, path, or cookie).
Selective Exposure: Include or exclude tags and operation IDs from the OpenAPI specification to control which endpoints are exposed.
Base URL Override: Manually override the API server base URL if needed.
Custom Naming/Description: Set default name and description for the generated MCP toolset.
Example Provided: Includes a Weatherbit API example, with Docker Compose setup for demonstration.
Command-Line Options: Multiple flags for fine-grained control over server behavior and toolset generation.

Installation

Docker (Recommended):
- Use the pre-built Docker image from Docker Hub or build locally.
- Run the container with required OpenAPI spec and API key configuration.
Docker Compose:
- Example setup provided in the repository for running with docker-compose.

Usage

Provide the OpenAPI specification (local file or remote URL).
Configure API keys as needed (via environment or .env file).
Run the server and connect MCP-compatible clients to the exposed APIs.
Use command-line flags to customize behavior (see features above).

Pricing

Open Source: No pricing information is provided; the project is available for free on GitHub.