AutoMCP 🚀

AutoMCP is a powerful platform that automatically generates ready-to-use Model Context Protocol (MCP) servers from your backend API documentation. Simply upload your OpenAPI, Swagger, or Postman collection, and AutoMCP uses Google's Gemini AI to generate a complete, production-ready MCP server package.

✨ Features

🤖 AI-Powered Generation: Uses Google Gemini AI to generate high-quality MCP server code
📚 Multiple Format Support: Supports OpenAPI 3.x, Swagger 2.x, and Postman Collections
🔧 Complete Package Generation: Generates full MCP server with:
- TypeScript tool implementations
- Type definitions
- Configuration files
- Package.json with dependencies
- Comprehensive README
🎯 Smart Parsing: Automatically extracts endpoints, parameters, schemas, and authentication schemes
📦 Ready-to-Use: Generated packages are immediately deployable
🔒 Secure: Local file storage, no cloud dependencies
🚀 Fast: Efficient parsing and code generation pipeline

📋 Prerequisites

Node.js >= 18.x
npm >= 9.x
Google Gemini API Key (Get one here)

🛠️ Installation

Clone the repository
```
git clone <repository-url>
cd AutoMCP
```
Install dependencies
```
npm install
```

Set up environment variables

cp .env.example .env

Edit .env and add your Gemini API key:

PORT=3000
GEMINI_API_KEY=your_gemini_api_key_here
GEMINI_MODEL=gemini-1.5-pro
GEMINI_MAX_TOKENS=32768
MAX_FILE_SIZE=10485760

Build the project
```
npm run build
```

⚙️ Configuration

Environment Variables

Variable	Description	Default
`PORT`	Server port	`3000`
`GEMINI_API_KEY`	Google Gemini API key	Required
`GEMINI_MODEL`	Gemini model to use	`gemini-1.5-pro`
`GEMINI_TEMPERATURE`	Generation temperature	`0.7`
`GEMINI_MAX_TOKENS`	Maximum tokens per request	`32768`
`MAX_FILE_SIZE`	Max upload file size (bytes)	`10485760` (10MB)
`NODE_ENV`	Environment mode	`development`

Supported Gemini Models

gemini-1.5-pro (Recommended)
gemini-pro
gemini-1.5-flash

To list available models:

npm run list-models

🚀 Usage

Start the Server

Development mode:

npm run dev

Production mode:

npm run build
npm start

The server will start on http://localhost:3000

API Workflow

1. Upload API Documentation

curl -X POST http://localhost:3000/api/upload \
  -F "file=@your-api-docs.json"

Response:

{
  "status": "success",
  "fileId": "8f4be01f-236d-4f67-8b70-6cf347b81d12",
  "filename": "pet-store-openapi.json",
  "metadata": {
    "title": "Swagger Petstore",
    "version": "1.0.0",
    "endpoints": 19
  }
}

2. Generate MCP Server

curl -X POST http://localhost:3000/api/generate \
  -H "Content-Type: application/json" \
  -d '{
    "fileId": "8f4be01f-236d-4f67-8b70-6cf347b81d12"
  }'

Response:

{
  "status": "success",
  "packageId": "8f4be01f-236d-4f67-8b70-6cf347b81d12",
  "packageName": "mcp-swagger-petstore---openapi-3-0",
  "downloadUrl": "/api/download/8f4be01f-236d-4f67-8b70-6cf347b81d12",
  "toolsGenerated": 19
}

3. Check Generation Status

curl http://localhost:3000/api/status/8f4be01f-236d-4f67-8b70-6cf347b81d12

4. Download Generated Package

curl -O http://localhost:3000/api/download/8f4be01f-236d-4f67-8b70-6cf347b81d12

Or open in browser:

http://localhost:3000/api/download/8f4be01f-236d-4f67-8b70-6cf347b81d12

Using Postman

Import the provided Postman collection (AutoMCP.postman_collection.json) for easy testing. See POSTMAN_SETUP.md for details.

📚 API Documentation

Endpoints

`GET /api/health`

Health check endpoint.

Response:

{
  "status": "ok",
  "timestamp": "2024-01-15T10:30:00.000Z",
  "uptime": 3600
}

`POST /api/upload`

Upload API documentation file (OpenAPI/Swagger/Postman).

Request:

Method: POST
Content-Type: multipart/form-data
Body: file (file field)

Response:

{
  "status": "success",
  "fileId": "uuid",
  "filename": "api-docs.json",
  "metadata": { ... }
}

`POST /api/generate`

Generate MCP server from uploaded file.

Request:

{
  "fileId": "uuid"
}

Response:

{
  "status": "success",
  "packageId": "uuid",
  "packageName": "mcp-api-name",
  "downloadUrl": "/api/download/uuid",
  "toolsGenerated": 19
}

`GET /api/download/:id`

Download generated MCP server package as ZIP.

Response:

Content-Type: application/zip
File: mcp-server-package.zip

`GET /api/status/:id`

Get generation status of a package.

Response:

{
  "status": "completed",
  "packageId": "uuid",
  "packageName": "mcp-api-name",
  "toolsGenerated": 19,
  "createdAt": "2024-01-15T10:30:00.000Z"
}

🏗️ Architecture

AutoMCP
├── API Layer (Express)
│   ├── Upload Controller
│   ├── Generate Controller
│   └── Download Controller
├── Parser Layer
│   ├── OpenAPI Parser
│   ├── Schema Resolver
│   └── Data Transformer
├── Gemini Integration
│   ├── Gemini Client (with retry logic)
│   ├── Code Generator
│   └── Prompt Templates
├── Generator Layer
│   ├── MCP Package Generator
│   └── Config Template Generator
└── Storage Layer
    ├── File Storage (Local)
    └── Package Storage (Local)

Data Flow

Upload → File saved to temp storage
Parse → OpenAPI/Swagger parsed into structured format
Transform → Data normalized for Gemini prompts
Generate → Gemini AI generates MCP server code
Assemble → Code assembled into complete package
Package → ZIP archive created
Download → Package available for download

🧪 Testing the Generated MCP Server

1. Extract and Install

unzip mcp-server-package.zip
cd mcp-server-package
npm install

2. Configure

Create .env file:

API_BASE_URL=https://your-api.com/v2
API_KEY=your_api_key

3. Build

npm run build

4. Test with MCP Inspector

npx @modelcontextprotocol/inspector node dist/index.js

This opens a web UI at http://localhost:5173 where you can test all generated tools.

5. Test Individual Tools

node -e "
import('./dist/tools/getPetById.js').then(m => 
  m.getPetById({petId: 1})
    .then(console.log)
    .catch(console.error)
);
"

🔧 Development

Project Structure

src/
├── api/              # API routes and controllers
├── config/           # Configuration management
├── gemini/           # Gemini AI integration
├── generators/       # MCP package generators
├── middleware/       # Express middleware
├── parsers/          # API documentation parsers
├── transformers/     # Data transformers
├── types/            # TypeScript type definitions
├── utils/            # Utility functions
└── server.ts         # Main server entry point

Development Commands

# Development with hot reload
npm run dev

# Build TypeScript
npm run build

# Production start
npm start

# List available Gemini models
npm run list-models

Code Style

TypeScript with strict mode
ES2020 target
CommonJS modules
Express.js for API
Async/await for async operations

🐛 Troubleshooting

Common Issues

1. Gemini API Errors

Error: 404 Not Found - models/gemini-X is not found

Solution:

Check available models: npm run list-models
Update GEMINI_MODEL in .env to a valid model
The system will auto-detect and use available models

2. File Upload Errors

Error: File too large

Solution:

Increase MAX_FILE_SIZE in .env
Or compress your API documentation file

3. Generation Timeout

Error: Generation timeout

Solution:

Increase GEMINI_MAX_TOKENS in .env
Split large APIs into smaller services
Check Gemini API quota limits

4. Generated MCP Server Errors

Error: API error: 500 Internal Server Error

Solution:

Check the API base URL in generated config.ts
Verify API authentication credentials
Test the API directly with curl

Debug Mode

Enable verbose logging:

DEBUG=* npm run dev

📝 Generated Package Structure

mcp-api-name/
├── src/
│   ├── index.ts          # MCP server entry point
│   ├── config.ts         # Configuration
│   ├── types.ts          # TypeScript types
│   └── tools/            # Generated tool implementations
│       ├── getPetById.ts
│       ├── addPet.ts
│       └── ...
├── package.json          # Dependencies
├── tsconfig.json         # TypeScript config
├── .env.example          # Environment template
├── .gitignore
└── README.md             # Usage instructions

🤝 Contributing

Contributions are welcome! Please follow these steps:

Fork the repository
Create a feature branch (git checkout -b feature/amazing-feature)
Commit your changes (git commit -m 'Add amazing feature')
Push to the branch (git push origin feature/amazing-feature)
Open a Pull Request

📄 License

ISC License

🙏 Acknowledgments

Model Context Protocol - The MCP specification
Google Gemini AI - AI code generation
OpenAPI Initiative - API specification standard

📞 Support

For issues, questions, or contributions:

Open an issue on GitHub
Check existing documentation
Review troubleshooting section

🗺️ Roadmap

Built with ❤️ for the hackathon

Name		Name	Last commit message	Last commit date
Latest commit History 5 Commits
public		public
src		src
.gitignore		.gitignore
AutoMCP.postman_collection.json		AutoMCP.postman_collection.json
FRONTEND.md		FRONTEND.md
GEMINI_API_UPDATES.md		GEMINI_API_UPDATES.md
IMPLEMENTATION_PLAN.md		IMPLEMENTATION_PLAN.md
POSTMAN_SETUP.md		POSTMAN_SETUP.md
PROJECT-DESCRIPTION.md		PROJECT-DESCRIPTION.md
README.md		README.md
nodemon.json		nodemon.json
package-lock.json		package-lock.json
package.json		package.json
pet-store-openapi.json		pet-store-openapi.json
pet-store-swagger.json		pet-store-swagger.json
tsconfig.json		tsconfig.json

pranavkale07/AutoMCP

Folders and files

Latest commit

History

Repository files navigation

AutoMCP 🚀

✨ Features

📋 Prerequisites

🛠️ Installation

⚙️ Configuration

Environment Variables

Supported Gemini Models

🚀 Usage

Start the Server

API Workflow

1. Upload API Documentation

2. Generate MCP Server

3. Check Generation Status

4. Download Generated Package

Using Postman

📚 API Documentation

Endpoints

GET /api/health

POST /api/upload

POST /api/generate

GET /api/download/:id

GET /api/status/:id

🏗️ Architecture

Data Flow

🧪 Testing the Generated MCP Server

1. Extract and Install

2. Configure

3. Build

4. Test with MCP Inspector

5. Test Individual Tools

🔧 Development

Project Structure

Development Commands

Code Style

🐛 Troubleshooting

Common Issues

1. Gemini API Errors

2. File Upload Errors

3. Generation Timeout

4. Generated MCP Server Errors

Debug Mode

📝 Generated Package Structure

🤝 Contributing

📄 License

🙏 Acknowledgments

📞 Support

🗺️ Roadmap

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Languages

`GET /api/health`

`POST /api/upload`

`POST /api/generate`

`GET /api/download/:id`

`GET /api/status/:id`

Packages