Overview
What is Apifox MCP Server?
Apifox MCP Server provides Apifox project API documentation as data to AI programming IDE tools like Cursor, enabling AI to directly access project API documentation.
How to use Apifox MCP Server?
Install Node.js 18+, generate an Apifox access token, obtain a project ID, then add the MCP configuration JSON to your IDE's MCP configuration file (e.g., ~/.cursor/mcp.json or Cline's MCP Server settings).
Key features of Apifox MCP Server
- Automatically reads all API docs from an entire Apifox project
- Caches API documentation data locally on your computer
- Provides API documentation data to AI via the Model Context Protocol
- Supports direct reading of Swagger/OAS files
- Supports multiple project API documentation configurations
Use cases of Apifox MCP Server
- Generate Product model definition code from API documentation
- Add new fields to a Product DTO based on updated API docs
- Add comments to each field in a Product class using API documentation
- Generate all MVC code related to a specific endpoint (e.g.,
/users)
FAQ from Apifox MCP Server
How does AI access the API documentation data?
AI reads all API documentation data from the entire Apifox project through the MCP protocol, with data cached locally on your computer.
How do I refresh API documentation data if Apifox data updates?
Tell AI to refresh the API documentation data, otherwise AI may read outdated cached data.
Which IDEs are supported?
Apifox MCP Server supports Cursor and VSCode with the Cline plugin, both of which support the MCP protocol.
What API documentation formats are supported?
Besides Apifox projects, the server also supports reading Swagger/OAS files directly via the --oas parameter (URL or local file path).
What should I note when configuring the MCP Server?
Name the MCP Server with "API文档" in its name for AI to recognize its purpose, avoid leaking access tokens in shared repositories, and add --apifox-api-base-url for on-premises deployment users.
