An AI-powered GitHub Repository Intelligence Assistant built using Large Language Models (LLMs), LangChain, FastAPI, FAISS, and Streamlit.
Understand any GitHub repository in minutes, not hours. 🏆 Built during the IBM Bob × lablab.ai Hackathon using LangChain, OpenAI GPT-4, FAISS, FastAPI, and Streamlit to simplify GitHub repository understanding through AI-powered analysis.
Built for the IBM Bob × lablab.ai Hackathon 🏆
Developers waste countless hours:
- 📚 Understanding unfamiliar codebases
- 🔍 Searching for specific functionality
- 📝 Writing documentation
- 👥 Onboarding new team members
- 🔎 Reviewing pull requests
Existing AI coding assistants generate code but lack deep repository understanding.
CodeOrbit AI is an intelligent assistant that:
✅ Analyzes any GitHub repository in minutes
✅ Explains architecture and code structure
✅ Answers questions about the codebase
✅ Generates documentation automatically
✅ Reviews pull requests for risks
✅ Guides new developers through onboarding
- Clone and parse any public GitHub repository
- Extract code structure, dependencies, and patterns
- Identify entry points and key files
- Generate comprehensive statistics
- Ask questions about the codebase in natural language
- Get context-aware answers with code examples
- Retrieve relevant code snippets automatically
- Maintain conversation history
- Auto-generate README files
- Create onboarding guides for new developers
- Generate API documentation
- Export in multiple formats (Markdown, PDF, HTML)
- Analyze pull requests for potential issues
- Detect security vulnerabilities
- Identify performance bottlenecks
- Suggest improvements
This project was developed as part of the IBM Bob × lablab.ai Hackathon.
The goal was to build an AI-powered solution that helps developers understand unfamiliar GitHub repositories through intelligent code analysis, repository chat, automated documentation generation, and developer onboarding.
The project demonstrates practical applications of Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), semantic search, and modern AI development tools.
┌─────────────────┐
│ Streamlit UI │ ← User Interface
└────────┬────────┘
│
▼
┌─────────────────┐
│ FastAPI Server │ ← Core Processing
└────────┬────────┘
│
┌────┴────┬──────────┬──────────┐
▼ ▼ ▼ ▼
┌────────┐ ┌──────┐ ┌────────┐ ┌──────────┐
│ GitHub │ │ FAISS│ │LangChain│ │ OpenAI │
│ API │ │ DB │ │ Agents │ │ API │
└────────┘ └──────┘ └────────┘ └──────────┘
Frontend:
- 🎨 Streamlit - Interactive web interface
- 📊 Plotly - Data visualization
Backend:
- ⚡ FastAPI - High-performance API server
- 🐍 Python 3.9+ - Core language
AI & ML:
- 🤖 LangChain - LLM orchestration
- 🧠 OpenAI GPT-4 - Language model
- 📚 FAISS - Vector similarity search
- 🔤 OpenAI Embeddings - Text embeddings
Repository Management:
- 📦 GitPython - Git operations
- 🌳 Tree-sitter - Code parsing
- Python 3.9 or higher
- OpenAI API key
- Git installed
✔ Codebase Search
✔ Semantic Search
✔ Embedding-based Retrieval
✔ Context-aware Chat
✔ Repository Statistics
✔ Architecture Analysis
✔ Intelligent Documentation
✔ Developer Onboarding
- Built a Retrieval-Augmented Generation (RAG) application.
- Worked with Large Language Models.
- Implemented semantic search using FAISS.
- Integrated FastAPI with Streamlit.
- Developed AI-powered developer tools.
- Improved understanding of repository intelligence systems.
Backend - https://code-orbit-ai.onrender.com/ Frontend - https://codeorbitai-aadj9flhdqzs2ntljrrr79.streamlit.app/
This project is licensed under the MIT License - see the LICENSE file for details.
- IBM Bob - For the amazing AI capabilities
- lablab.ai - For organizing the hackathon
- OpenAI - For GPT-4 and embeddings
- FastAPI - For the excellent framework
- Streamlit - For rapid UI development
- LangChain - For LLM orchestration
-
Email: harmaienmalika@gmail.com
-
LinkedIn: linkedin.com/in/malika-harmaien
Made with 🚀 for developers, by developers