Universal Deep Research (UDR)

A prototype research system that combines user-defined strategies, intelligent web search, content analysis, and automated report generation using large language models. This repository contains both the backend API service and frontend web interface.

As mentioned above, this is a research demonstration prototype and should not be used for production purposes. The software contains experimental features and research-grade implementations.

Project Structure

/
├── backend/          # FastAPI backend service
│   ├── README.md     # Backend setup and configuration
│   ├── main.py       # FastAPI application
│   ├── scan_research.py  # Core research logic
│   ├── frame/        # Advanced reliability framework
│   └── ...
├── frontend/         # Next.js frontend application
│   ├── README.md     # Frontend setup and configuration
│   ├── src/          # React components and pages
│   └── ...
└── README.md         # This file

Getting Started

To run the prototype, you need to start both the backend and frontend services:

1. Backend Setup

Navigate to the backend directory and follow the setup instructions:

cd backend

See backend/README.md for detailed setup instructions, including:

• Python environment setup
• API key configuration
• Server startup commands

2. Frontend Setup

In a new terminal, navigate to the frontend directory:

cd frontend

See frontend/README.md for detailed setup instructions, including:

• Node.js dependencies installation
• Environment configuration
• Development server startup

3. Running the Prototype

1. Start the backend server (typically on port 8000):

   cd backend
   launch_server.sh
   

2. Start the frontend development server (typically on port 3000):

   cd frontend
   npm run dev
   

3. Access the application: Open your browser and navigate to http://localhost:3000 or similar as output by the terminal.

Features

• Intelligent Research: User-configurable research strategies
• Real-time Progress: Live updates during research and report generation
• Interactive Interface: Modern web UI for research queries and results
• Multi-Model Support: Configurable LLM backends

Documentation

• Backend Documentation - API setup, configuration, and endpoints
• Frontend Documentation - UI setup, configuration, and deployment

Requirements

• Python 3.8+ (for backend)
• Node.js 18+ (for frontend)
• API keys for LLM providers (e.g., NVIDIA NGC, OpenAI, etc.)
• Tavily API key for web search functionality

Development

This is a research prototype demonstrating AI-powered research automation concepts. The codebase is structured for experimentation and demonstration rather than production deployment.