🚀 RepoSense AI

LIVE DEMO - https://reposense-ai-hcpmx9syan2oz2r7uri2ad.streamlit.app/

AI-Powered Repository Analysis & Developer Onboarding Platform

RepoSense AI is an AI-powered developer assistant that analyzes GitHub repositories using Google Gemini, LangChain, FAISS, and Retrieval-Augmented Generation (RAG). It helps developers understand unfamiliar codebases within minutes by generating repository summaries, answering repository-specific questions, creating onboarding guides, performing impact analysis, and automatically generating professional README files.

---

❗ Problem Statement

Understanding an unfamiliar codebase is one of the biggest challenges for developers.

When joining a new project, developers often spend hours or even days:

• 📂 Exploring hundreds of source files
• 🔍 Searching for the application's entry point
• 🧩 Understanding module relationships
• 📖 Reading incomplete or outdated documentation
• ⚠️ Predicting the impact of code changes
• 🤝 Onboarding without proper guidance

Traditional documentation quickly becomes outdated, making repository exploration slow and inefficient.

---

💡 Proposed Solution

RepoSense AI leverages Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and repository analysis to automatically understand a GitHub repository and provide intelligent insights.

The platform enables developers to:

• 🤖 Generate architecture-aware repository summaries
• 💬 Chat with the repository using natural language
• 🌱 Create developer onboarding guides
• ⚡ Perform impact analysis before modifying code
• 📝 Generate professional README documentation
• 📊 Detect technologies and project structure automatically

---

🏗️ Solution Overview

RepoSense AI performs the following workflow:

1. 📥 Clone the GitHub repository
2. 📂 Parse the project structure
3. 🔍 Detect technologies, frameworks, databases, and entry points
4. 📄 Extract source code and documentation
5. 🧠 Generate embeddings using Sentence Transformers
6. 📚 Store embeddings in a FAISS Vector Database
7. 🤖 Query Google Gemini using LangChain RAG
8. 📊 Present repository insights through an interactive Streamlit dashboard

---

✨ Features

📂 Repository Analysis

• Clone any public GitHub repository
• Detect:
  • Programming Language
  • Framework
  • Database
  • Package Manager
  • Build Tool
  • Entry Point
  • Repository Complexity
  • Folder Hierarchy

---

🤖 AI Repository Summary

Automatically generates:

• Project Purpose
• Technology Stack
• Overall Architecture
• Folder Structure
• Entry Point
• Main Modules
• Request Flow
• Important Files
• Suggested Reading Order

---

💬 Repository Chat (RAG)

Ask questions about any repository using natural language.

---

🌱 Developer Onboarding

Generate onboarding documentation for new contributors.

Includes:

• Learning Roadmap
• Module Explanation
• Development Workflow
• Reading Sequence
• Important Files

---

⚡ Impact Analysis

Predict the consequences of changing any file before modifying it.

---

📝 README Generator

Automatically generates a professional README.md for the repository.

---

💬 Example Repository Chat Questions

• Explain this repository.
• What is the overall architecture?
• Which framework is being used?
• Explain the backend folder.
• Where is the application's entry point?
• How does authentication work?
• Which files are responsible for API calls?
• Which file should I read first?
• Explain the request flow.
• Summarize the repository in simple language.

---

⚡ Example Impact Analysis Questions

• What happens if I modify repository_service.py?
• Which files depend on this module?
• What components will be affected?
• Which APIs will break?
• What should I regression test?
• Suggest a safe refactoring strategy.
• What are the downstream effects of changing this class?
• Which modules import this file?
• Will changing this function affect other components?
• Estimate the risk of modifying this file.

---

🛠️ Tech Stack

Category	Technology
💻 Language	Python
🎨 Frontend	Streamlit
🤖 LLM	Google Gemini 2.5 Flash
🔗 AI Framework	LangChain
📚 Vector Database	FAISS
🧠 Embeddings	Sentence Transformers
📦 Git Operations	GitPython
📄 Environment	Python Dotenv

---

📂 Project Structure

RepoSense-AI
│
├── api/
│   ├── chat_api.py
│   ├── impact_api.py
│   ├── onboarding_api.py
│   ├── repository_api.py
│   └── summary_api.py
│
├── backend/
│   ├── embedding_service.py
│   ├── impact_service.py
│   ├── onboarding_service.py
│   ├── parser.py
│   ├── rag_service.py
│   ├── repository_service.py
│   ├── summary_service.py
│   ├── utils.py
│   └── faiss_index/
│
├── frontend/
│   ├── components/
│   ├── pages/
│   └── app.py
│
├── repositories/
├── main.py
├── requirements.txt
└── README.md

---

🚀 Future Scope

• 🔒 Private Repository Support
• 🐳 Docker Deployment
• ☁️ Cloud Deployment
• 📈 Dependency Graph Visualization
• 🧪 Automatic Test Generation
• 🔍 Multi-Repository Comparison
• 📊 Code Quality Dashboard
• 🤝 GitHub Pull Request Review Assistant
• 📝 Automatic Documentation Updates
• ⚙️ CI/CD Integration

---

⚙️ Clone & Run

git clone https://github.com/Chirag04-bit/RepoSense-AI.git

cd RepoSense-AI

python -m venv .venv

# Windows
.venv\Scripts\activate

# Linux/Mac
source .venv/bin/activate

pip install -r requirements.txt

Create a .env file:

GOOGLE_API_KEY=YOUR_GEMINI_API_KEY

Run the application:

streamlit run main.py

---