Repo-IQ 🚀

An AI-powered GitHub Repository exploration tool built using a RAG (Retrieval-Augmented Generation) pipeline. This project dynamically clones a GitHub repository, processes its source files into vector embeddings using ChromaDB, and allows you to chat with the codebase using Google's Gemini AI.

---

✨ Features

• Dynamic Ingestion: Pass any public GitHub URL and the pipeline automatically clones, chunks, and vectorizes the codebase.
• Context-Aware AI Answers: Ask complex questions about the architecture, logic, or flow of the codebase. The Gemini model will answer using only the context retrieved from the source code.
• Memory/History Aware: The AI remembers your previous questions in the session, allowing for conversational follow-ups.
• Slate-Dark UI: A stunning, modern, and responsive dark interface built with React and Vite that communicates seamlessly with the Flask backend.
• Instant Status Feedback: Clear visibility into clone status and success feedback directly integrated into the dashboard.

---

🛠️ Tech Stack

• Backend: Python, Flask, Flask-CORS, LangChain, ChromaDB, Gunicorn
• AI/LLM: Google Gemini 2.5 Flash (langchain-google-genai)
• Frontend: React (Vite), React Router DOM, Axios, CSS3

---

⚙️ Installation & Setup

1. Clone or Navigate to the Project Directory

cd Repo-IQ

2. Backend Setup

1. Navigate to the backend folder:

   cd backend

2. Install Python dependencies:

   pip install -r requirements.txt

3. Set up the Environment Variables: Create or open a .env file in the project root directory (or in the backend/ directory depending on your runner configuration) and configure your Google Gemini API Key:

   GOOGLE_API_KEY=your_actual_api_key_here

3. Frontend Setup

1. Navigate to the client folder:

   cd ../client

2. Install Node dependencies:

   npm install

---

🚀 How to Run

1. Start the Backend Server

From the backend directory, run:

python app.py

The server will start on http://127.0.0.1:5000.

2. Start the Frontend Dev Server

From the client directory, run:

npm run dev

Vite will start the frontend local server (typically at http://localhost:5173). Open this URL in your browser.

3. Analyze a Repository

1. Paste a public GitHub URL into the input field (e.g., https://github.com/username/project).
2. Click Analyze.
3. Wait for cloning and vector embedding to complete (you'll see an "Analyse complete, you can ask questions about the repo" green message).
4. Start chatting with the codebase!

---

Made with ❤️ by Team: Chicken Biryani (Deon George, Karthik Krishnan, and Dhanush M) In Association with BeyondPrompt-26 by IEEE CS GECT