next-mcp-server

MCP.Pizza Chef: vertile-ai

The next-mcp-server is an MCP server designed to analyze Next.js applications by scanning their directory structure to extract comprehensive API route information. It provides details such as API paths, HTTP methods, request parameters, status codes, and request/response schemas. This enables LLMs to better understand and interact with Next.js apps, facilitating advanced AI-driven workflows and integrations.

Use This MCP server To

Extract API route metadata from Next.js projects Provide LLMs with structured Next.js API route context Enable AI agents to generate API documentation automatically Support automated testing by exposing API schemas Facilitate code generation based on Next.js API structure Integrate Next.js API details into AI-powered developer tools

README

Next.js MCP Server

Demo

Router analysis demo

Features

  • get-routers-info

    The Router Analyzer scans your Next.js app directory structure and extracts information about all API routes, including:

    • API paths
    • HTTP methods (GET, POST, PUT, DELETE, etc.)
    • Request parameters
    • Status codes
    • Request and response schemas

Installation

npm install next-mcp-server

Or if you're using pnpm:

pnpm add next-mcp-server

Usage

Command Line

You can run the mcp server directly:

npm run build
node dist/index.js

Docker

docker build -t mcp/next -f Dockerfile .
docker run mcp/next -d

For cursor usage, define a mcp.json under ~/.cursor or [projectDir]/.cursor

{
  "mcpServers": {
    "next.js": {
      "url": "http://localhost:4857/sse"
    }
  }
}

The url here could vary based on your .env settings within the project.

Output

The tool generates detailed information about each route:

[
  {
    "filePath": "/path/to/your/app/api/test/route.ts",
    "implementationPath": "/path/to/your/app/api/test/route.ts",
    "apiPath": "/api/test",
    "handlers": [
      {
        "method": "GET",
        "path": "/api/test",
        "functionSignature": "export async function GET(request: Request)",
        "description": "Get test data",
        "parameters": [],
        "statusCodes": [200]
      },
      {
        "method": "POST",
        "path": "/api/test",
        "functionSignature": "export async function POST(request: Request)",
        "description": "Create test data",
        "parameters": [],
        "requestBodySchema": "{ name: string }",
        "statusCodes": [201, 400]
      }
    ]
  }
]

Development

To run tests:

npm run test

To run the mcp server locally:

npm run build
node dist/index.js 

To run it from node_modules after npm i:

node node_modules/next-mcp-server/dist/index.js

To run the playground:

pnpm --filter playground dev

How It Works

The tool:

  1. Scans your Next.js app directory structure for route files
  2. Analyzes each route file to extract HTTP methods, paths, parameters, etc.
  3. Extracts documentation from comments
  4. Returns a structured representation of all your API routes

Restrictions

  1. Due to the nature of accessing filesystem directory by path, it will not work if hosted over network
  2. Only supports Next.js App router projects

License

MIT

next-mcp-server FAQ

How do I install next-mcp-server?
Install it via npm with 'npm install next-mcp-server' or pnpm with 'pnpm add next-mcp-server'.
How do I run the next-mcp-server?
You can run it directly using 'npm run build' followed by 'node dist/index.js', or use the provided Docker image.
What information does next-mcp-server extract from Next.js apps?
It extracts API paths, HTTP methods, request parameters, status codes, and request/response schemas.
How do I configure next-mcp-server for use with cursor?
Define a 'mcp.json' file under '~/.cursor' or your project directory with the server URL under 'mcpServers.next.js.url'.
Can next-mcp-server handle custom environment configurations?
Yes, the server URL can vary based on your project's .env settings to accommodate different environments.
What benefits does next-mcp-server provide for AI models?
It supplies structured, detailed API context enabling LLMs like OpenAI, Claude, and Gemini to better understand Next.js APIs.
Is next-mcp-server limited to API routes only?
Currently, it focuses on API routes, including their methods and schemas, to provide rich context for LLMs.