Frequently Asked Questions (FAQ) about Swagger MCP Server

Q: What is a Swagger MCP Server? A: An MCP server that connects to a Swagger/OpenAPI definition and helps an AI build all the required models to generate an MCP server for that service.

Q: What are the key features of the Swagger MCP Server? A: Key features include downloading Swagger specifications, listing endpoints and models, generating model code, and creating MCP tool definitions with AI-specific instructions.

Q: What are the prerequisites for using the Swagger MCP Server? A: You need Node.js (v14 or higher) and npm or yarn installed.

Q: How do I install the Swagger MCP Server? A: Clone the repository, install dependencies using npm install, and configure the .env file.

Q: How do I run the Swagger MCP Server? A: Build the application with npm run build and then run it using node build/index.js.

Q: How do I add the Swagger MCP Server to Cursor? A: In Cursor settings, add a new MCP server, select “stdio” as the transport type, and enter the command to run the server: node path/to/swagger-mcp/build/index.js.

Q: What tools are available through the Swagger MCP Server? A: Available tools include getSwaggerDefinition, listEndpoints, listEndpointModels, generateModelCode, and generateEndpointToolCode.

Q: What are MCP prompts, and how are they used? A: MCP prompts guide AI assistants through common workflows by providing step-by-step instructions for tasks like adding new endpoints. Clients can make a prompts/get request with the prompt name and arguments.

Q: How does the Swagger MCP Server associate with a specific Swagger API? A: After downloading a Swagger file, the server creates a .swagger-mcp configuration file that stores the filename of the locally stored Swagger file.

Q: What improvements have been made to the MCP tool code generator? A: Improvements include complete schema information, better parameter naming, semantic tool names, support for YAML Swagger files, improved documentation, no external dependencies, and AI-specific instructions.

Q: How do I generate an MCP tool definition for an endpoint? A: Use the generateEndpointToolCode function, providing the path, method, Swagger file path, and singularizeResourceNames option.

Q: Where can I find the collection of MCP prompts? A: Check the PROMPTS.md file for the full collection of prompts.

Q: How does UBOS leverage MCP Server? A: UBOS platform leverages MCP to create robust and context-aware AI Agents.