AI-NAD: Autonomous AI Software Factory

AI-NAD converts natural language intent into runnable applications using a multi-agent AI pipeline.

Features

• Multi-Agent Architecture: Intent, Requirement, Architecture, Code, Test, and Validation agents
• Local AI Models: Uses Ollama for free, local AI model execution
• Full Stack Generation: Generates complete React frontend and Express backend
• Automatic Testing: Generates unit tests for generated code
• Code Validation: Validates generated code for syntax, security, and dependencies

Prerequisites

1. Node.js (v18 or higher)

2. Gemini API Key (recommended) or Ollama installed and running

   Option 1: Gemini API (Recommended)

   • Get your API key from Google AI Studio
   • No local installation required

   Option 2: Ollama (Local)

   # Install Ollama from https://ollama.ai
   # Pull a model (e.g., deepseek-coder)
   ollama pull deepseek-coder

Installation

1. Clone the repository

2. Install dependencies:

   npm run install:all

3. Configure environment variables:

   cd backend
   cp .env.example .env
   # Edit .env with your API configuration
   # For Gemini: Set GEMINI_API_KEY=your-api-key
   # For Ollama: Set AI_SERVICE_TYPE=ollama and ensure Ollama is running

Running the Application

Development Mode

1. Start the backend:

   npm run dev:backend

2. In another terminal, start the frontend:

   npm run dev:frontend

3. Open http://localhost:3000 in your browser

Usage

1. Enter a project description in the web interface

   • Example: "Create a task manager web app with login and dashboard"

2. Click "Generate Project"

3. Wait for the AI agents to process your request

4. Once complete, navigate to the generated project in generated-projects/

5. Follow the README in the generated project to run it

Architecture

Agents

• Intent Agent: Converts natural language to structured requirements
• Requirement Agent: Breaks down requirements into services and entities
• Architecture Agent: Designs system architecture and API structure
• Code Agent: Generates production-ready code
• Test Agent: Creates unit and integration tests
• Validation Agent: Validates code quality and security

Pipeline Flow

User Prompt
   ↓
Intent Agent
   ↓
Requirement Agent
   ↓
Architecture Agent
   ↓
Code Agent
   ↓
Test Agent
   ↓
Validation Agent
   ↓
Project Builder
   ↓
Generated Project

Technology Stack

• Backend: Node.js, TypeScript, Express
• Frontend: React, Vite, TypeScript
• AI: Ollama (local models)
• Database: PostgreSQL (for generated projects)

Configuration

Edit backend/.env to configure:

For Gemini API (Recommended):

• AI_SERVICE_TYPE: Set to gemini
• GEMINI_API_KEY: Your Gemini API key from Google AI Studio
• GEMINI_MODEL: Model to use (default: gemini-1.5-flash)

For Ollama (Local):

• AI_SERVICE_TYPE: Set to ollama
• OLLAMA_BASE_URL: Ollama API URL (default: http://localhost:11434)
• OLLAMA_MODEL: Model to use (default: deepseek-coder)

General:

• PORT: Backend server port (default: 5000)

Generated Projects

All generated projects are stored in generated-projects/ directory. Each project includes:

• Complete backend with Express API
• React frontend with Vite
• Database schema
• Tests
• README with setup instructions

Security

The system includes:

• Prompt injection protection
• Code security validation
• Dependency vulnerability checks

License

MIT