🧭 CodeAtlas AI

Understand any GitHub repository in minutes, not days.

🎯 Impact

Input: GitHub repository URL
Output: Tech stack + AI summary + Architecture diagram + Engineering recommendations
Time to Insight: Under 60 seconds

---

🏆 Microsoft Build AI Hackathon 2026

Team Name: CodeAtlas

---

🔗 Live Demo & Resources

Resource	Link
🚀 Live Frontend	https://codeatlas-frontend-khaki.vercel.app/
⚙️ Backend API	https://codeatlas-backend-5hnr.onrender.com/docs
💻 Frontend Repository	https://github.com/sejalsksagar/codeatlas-frontend/
🛠️ Backend Repository	https://github.com/sejalsksagar/codeatlas-backend/

🎬 Theme

AI at Work: Productivity & Teamwork Reimagined

🔍 Problem Statement

Developers spend significant time understanding unfamiliar codebases.

⚡ Common challenges:

• 📚 Missing or outdated documentation
• 🖼️ No architecture diagrams
• 👨‍💼 Knowledge concentrated with senior engineers
• 🎓 Slow onboarding for new contributors
• 📦 Difficulty evaluating open-source projects

These challenges reduce developer productivity and make collaboration harder across teams.

💡 Solution

CodeAtlas AI transforms any public GitHub repository into understandable engineering knowledge.

Paste a GitHub repository URL and receive:

• 🔧 Technology stack detection
• 📝 Project summary in plain English
• 📦 Module-level explanations
• 🏗️ Architecture diagram visualization
• 🔒 Security recommendations
• ⚡ Performance recommendations
• 📈 Scalability recommendations

Our goal is to reduce repository onboarding from days to minutes.

🏗️ Architecture Overview

👤 User
  ↓
🎨 Next.js Frontend (Vercel)
  ↓
⚙️ FastAPI Backend (Render)
  ↓
🐙 GitHub API
  ↓
📦 Repository Structure Analysis
  ↓
🧠 GitHub Models (GPT-5)
  ↓
📝 Mermaid Architecture Generation
  ↓
🎭 React Flow Visualization
  ↓
📊 Summary + Diagram + Suggestions

🛠️ Technology Stack

🔧 Backend

• FastAPI
• Pydantic
• Python

🎨 Frontend

• Next.js 14
• React
• TypeScript
• Tailwind CSS
• React Flow

🤖 AI Layer

• GitHub Models (GPT-5)
• GitHub REST API

☁️ Infrastructure

• Render
• Vercel
• GitHub Actions

---

📸 Screenshots

1. Landing Page

Repository URL input screen where users can submit any public GitHub repository for analysis.

Highlights:

• Clean onboarding experience
• Repository URL validation
• Single-click analysis workflow

---

2. Repository Analysis Dashboard

Generated repository overview showing detected technology stack, AI-generated project summary, module explanations, and request flow.

Highlights:

• Automatic tech stack detection
• Plain-English repository understanding
• Module-level insights
• AI-powered analysis using GitHub Models GPT-5

---

3. Architecture Diagram Visualization

GPT-5 generated architecture diagram rendered using React Flow.

Highlights:

• Automatic architecture generation
• Interactive visualization
• Drag-and-reposition nodes
• Clear system understanding at a glance

---

4. Engineering Recommendations

AI-generated recommendations categorized by security, performance, scalability, and engineering quality.

Highlights:

• Actionable repository insights
• Security recommendations
• Performance improvement opportunities
• Scalability suggestions

🧩 Microsoft AI Stack Usage

This project leverages the Microsoft ecosystem through:

• 🧠 GitHub Models (GPT-5)
• 🐙 GitHub REST API
• ⚙️ GitHub Actions
• 🤖 GitHub Copilot

GitHub Models GPT-5 powers repository understanding, architecture generation, and engineering recommendations.

🤖 AI Tools Disclosure

The following AI tools were used during development:

• 💬 ChatGPT – architecture discussions and implementation guidance
• 🧠 Claude – planning and code review assistance
• ✨ Gemini – alternative solution exploration
• 🤖 GitHub Copilot – README generation, documentation assistance, and developer productivity support
• 🧠 GitHub Models GPT-5 – runtime AI inference inside the application

All generated code and outputs were reviewed, tested, and validated by the team.

🚀 Setup Instructions

Backend

git clone https://github.com/sejalsksagar/codeatlas-backend.git
cd codeatlas-backend
python -m venv venv
pip install -r requirements.txt
uvicorn main:app --reload

Environment Variables:

GITHUB_TOKEN=
GITHUB_MODELS_TOKEN=
ALLOWED_ORIGINS=

Frontend

For frontend setup, refer to:

https://github.com/sejalsksagar/codeatlas-frontend

🔄 CI/CD

GitHub Actions automatically executes backend tests on every push.

🚀 Future Improvements

• ✏️ Fully editable React Flow diagrams
• 📥 Diagram export functionality
• 💬 Repository chat assistant
• 📊 UML diagram generation
• 🗄️ Database schema visualization
• 🔒 Private repository support
• 🔀 Multi-repository analysis
• 📈 API usage analytics
• 🚀 Automated deployment workflows

👥 Team CodeAtlas

Sejal Kshirsagar

• Role: Backend Engineering & AI Integration
• LinkedIn: https://www.linkedin.com/in/sejalsksagar/

Priya Sharma

• Role: Frontend Engineering & UI/UX
• LinkedIn: https://www.linkedin.com/in/priya-sharma-mitwpu/

Soumyashri Singha

• Role: Architecture, Integration & Product Presentation
• LinkedIn: https://www.linkedin.com/in/soumyashri-singha-6850a2210/

🔗 Project Links

• 🚀 Frontend: https://codeatlas-frontend-khaki.vercel.app/
• ⚙️ Backend API: https://codeatlas-backend-5hnr.onrender.com/docs
• 💻 Frontend Repository: https://github.com/sejalsksagar/codeatlas-frontend/
• 🛠️ Backend Repository: https://github.com/sejalsksagar/codeatlas-backend/

---

Made with ❤️ by Team CodeAtlas for Microsoft Build AI Hackathon 2026