Advanced AI Research Assistant 🚀

A powerful, intelligent workspace that combines document notes with an autonomous, web-connected research agent.

---

📖 Overview

The Advanced AI Research Assistant transforms traditional note-taking into an autonomous intelligence hub. Powered by LangGraph and Google Generative AI (Gemini), the platform offers a powerful dual-model system: you can use the Local RAG Chatbot to strictly query your private uploaded documents, or engage the Autonomous Agent to perform live web searches, scrape web pages, and automatically dispatch comprehensive executive reports via email.

✨ Key Features

Feature	Description
🔍 Autonomous Web Searching	The agent dynamically breaks down your research topics and queries the open web (DuckDuckGo Search) to gather up-to-date intel.
🕸️ Deep Web Scraping	Uses web scraping to extract full text context from URLs, incorporating detailed external content directly into your reports.
📧 Automated Email Reports	Generates highly structured, executive summaries and automatically emails the final research report directly to your inbox.
📄 Document Processing & RAG	Upload PDFs and text directly into project "Notebooks." The system automatically embeds and indexes them for high-speed semantic search.
💬 Conversational RAG Chatbot	A dedicated, local chat interface that answers questions strictly using your uploaded documents, ensuring responses are grounded directly in your private context via FAISS.
🤖 Advanced LangGraph Agent	A floating, smart conversational assistant capable of routing between local document tools and web tools seamlessly.
📝 Strict Plain-Text Outputs	The agent interface ensures distraction-free, pure plain-text output, natively removing markdown for streamlined readability.
📓 Intelligent Notebook Management	Create distinct notebooks to elegantly organize personal documents, securely compartmentalized by user authentication.
🧹 Resource Cleanup	Safe, automated removal of related files and vector embeddings to keep your storage clean when a notebook is deleted.

🛠️ Technology Stack

Our platform leverages a robust and modern stack:

• Backend Engineering: Python 3, Flask, Werkzeug
• AI & Orchestration: LangChain, LangGraph, Google Generative AI (Gemini SDK)
• Vector Database (RAG): FAISS / ChromaDB for ultra-fast similarity search
• Web Toolkit: DuckDuckGo Search API, Requests, BeautifulSoup4
• Communication: SMTP Email Integration
• Database: SQLite for user and metadata management
• Frontend Design: HTML5, CSS3, Vanilla JavaScript, Jinja2 Templates

🚀 Getting Started

Follow these instructions to get a copy of the project up and running on your local machine for development and testing purposes.

Prerequisites

Ensure you have the following installed:

• Python 3.8 or higher
• Google Gemini API Key

Installation Steps

1. Clone the repository

   git clone <your-repository-url>
   cd GenAI

2. Set up the virtual environment

   python -m venv .venv
   
   # On Windows
   .venv\Scripts\activate
   
   # On macOS/Linux
   source .venv/bin/activate

3. Install dependencies

   pip install -r requirements.txt

4. Configure Environment Variables Create a .env file in the root directory and add your secret keys:

   # Core Config
   FLASK_SECRET_KEY="your_super_secret_key"
   
   # AI Config
   GOOGLE_API_KEY="your_google_gemini_api_key"
   
   # Email Config (for Report Dispatch)
   SMTP_SERVER="smtp.gmail.com"
   SMTP_PORT="587"
   EMAIL_USER="your_email@gmail.com"
   EMAIL_PASSWORD="your_app_password"

5. Run the Application

   python app.py

   🌐 The application will be accessible at http://localhost:5000.

📂 Project Structure

A quick look at the top-level files and directories you'll see in this project:

GenAI/
├── app.py                 # Main Flask application and API routing
├── requirements.txt       # Project dependencies
├── .env                   # Environment variables (API keys, secrets, SMTP)
├── database/              # Relational database models and initialization
├── rag/                   # AI logic, document processing, and RAG pipelines
│   ├── document_processor.py # Parses and chunks files
│   ├── faiss_store.py     # Vector database initialization and queries
│   ├── tools.py           # Web search, scraping, file ops, and email tools
│   └── agent.py           # LangGraph-based conversational agent
├── templates/             # Jinja2 HTML templates
├── static/                # Static assets (CSS, JS, Images)
└── uploads/               # Local directory for uploaded user documents

🤝 Contributing

Contributions, issues, and feature requests are always welcome!
Feel free to check out the issues page to report bugs or suggest enhancements.

📝 License

This project is generously licensed under the MIT License.