TAV Backend API

A Node.js backend API that integrates Google Gemini and OpenAI ChatGPT APIs. The API receives a prompt, sends it to Gemini, then combines the response with a configurable prepend prompt and sends it to ChatGPT for final processing.

Features

• Bearer Token Authentication: Protected endpoints with secure token validation
• Input Validation: Zod schema validation for all requests
• Rate Limiting: Built-in rate limiting to prevent abuse
• Security: Helmet.js for security headers and CORS support
• Error Handling: Comprehensive error handling and logging
• Environment Configuration: Flexible configuration via environment variables

API Flow

1. Receive POST request with prompt (protected by bearer token)
2. Send prompt to Google Gemini API
3. Take Gemini's response and prepend a configurable prompt
4. Send combined prompt/response to OpenAI ChatGPT API
5. Return ChatGPT's final response to the client

Setup

Prerequisites

• Node.js (v16 or higher) OR Docker
• npm or yarn (if not using Docker)
• Google Gemini API key
• OpenAI API key

Installation

1. Clone the repository:

git clone <repository-url>
cd tav-backend

2. Install dependencies:

npm install

3. Copy the environment example file:

cp env.example .env

4. Configure your environment variables in .env:

# Server Configuration
PORT=3000
NODE_ENV=development

# Authentication
BEARER_TOKEN=your-secure-bearer-token-here

# Google Gemini API
GEMINI_API_KEY=your-gemini-api-key-here
GEMINI_MODEL=gemini-pro

# OpenAI ChatGPT API
OPENAI_API_KEY=your-openai-api-key-here
OPENAI_MODEL=gpt-3.5-turbo

# Configuration
PREPEND_PROMPT=Please analyze and improve the following response:

# API Configuration
OPENAI_MAX_TOKENS=2000
OPENAI_TEMPERATURE=0.7

Running the Application

Option 1: Using Node.js directly

Development mode:

npm run dev

Production mode:

npm start

Option 2: Using Docker (Recommended)

Production with Docker:

docker-compose up --build

Development with Docker (hot reloading):

docker-compose -f docker-compose.dev.yml up --build

The server will start on port 3000 (or the port specified in your .env file).

For detailed Docker instructions, see DOCKER.md.

For VPS deployment instructions, see VPS_DEPLOYMENT.md.

API Documentation

Health Check

GET /health

Returns server status.

Response:

{
  "status": "OK",
  "timestamp": "2024-01-01T12:00:00.000Z"
}

Process Prompt

POST /api/prompt

Process a prompt through Gemini and ChatGPT.

Headers:

Authorization: Bearer your-bearer-token
Content-Type: application/json

Request Body:

{
  "prompt": "Your prompt here"
}

Response:

{
  "response": "ChatGPT's final response",
  "geminiResponse": "Gemini's original response",
  "processingTime": 1500
}

Error Responses:

• 400 Bad Request: Invalid request body or validation errors
• 401 Unauthorized: Missing or invalid bearer token
• 429 Too Many Requests: Rate limit exceeded
• 500 Internal Server Error: Server or API errors

Environment Variables

Variable	Description	Required	Default
PORT	Server port	No	3000
NODE_ENV	Environment (development/production)	No	development
BEARER_TOKEN	Authentication token	Yes	-
GEMINI_API_KEY	Google Gemini API key	Yes	-
GEMINI_MODEL	Gemini model to use	No	gemini-pro
OPENAI_API_KEY	OpenAI API key	Yes	-
OPENAI_MODEL	OpenAI model to use	No	gpt-3.5-turbo
OPENAI_MAX_TOKENS	Maximum tokens for OpenAI responses	No	2000
OPENAI_TEMPERATURE	Temperature for OpenAI responses (0-2)	No	0.7
PREPEND_PROMPT	Text to prepend to Gemini response	No	"Please analyze and improve the following response:"

Security Features

• Bearer Token Authentication: All API endpoints are protected
• Rate Limiting: 100 requests per 15 minutes per IP
• Security Headers: Helmet.js for security headers
• CORS: Configurable CORS support
• Input Validation: Zod schema validation
• Error Sanitization: Errors don't expose sensitive information in production

Error Handling

The API includes comprehensive error handling:

• Input validation errors (400)
• Authentication errors (401)
• Rate limiting (429)
• API service errors (500)
• Generic error responses with appropriate status codes

Development

Project Structure

src/
├── index.js          # Main application entry point
├── middleware/
│   └── auth.js       # Authentication middleware
├── routes/
│   └── prompt.js     # Prompt processing routes
├── schemas/
│   └── prompt.js     # Zod validation schemas
├── services/
│   ├── gemini.js     # Gemini API service
│   └── chatgpt.js    # ChatGPT API service
└── utils/
    └── config.js     # Configuration validation

# Docker files
Dockerfile            # Production Docker image
Dockerfile.dev        # Development Docker image
docker-compose.yml    # Production Docker Compose
docker-compose.dev.yml # Development Docker Compose
.dockerignore         # Docker ignore rules

Testing

Run tests:

npm test

Logs

The application logs important events:

• API requests and responses
• Error details
• Processing times
• Authentication attempts

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

License

MIT License