GitInsight AI

GitInsight AI is a simple GitHub profile analyzer built with FastAPI, Streamlit, and LangGraph.

Enter a GitHub username and the app reviews the profile using multiple AI-assisted agents, then shows:

• Profile overview
• Repository insights
• README and contribution signals
• Collaboration activity
• Recruiter readiness score
• AI mentor feedback

How It Works

The app uses an 8-agent workflow:

1. Profile Extractor
2. Repository Analyzer
3. README Evaluator
4. Contribution Intelligence
5. Code Quality Analyzer
6. Open Source Collaboration
7. Recruiter Readiness
8. AI Mentor

The flow is not fully linear:

• Profile Extractor runs first
• Repository Analyzer collects repo data
• The analysis then fans out to README, contribution, code-quality, and collaboration agents
• Their outputs are combined by Recruiter Readiness
• AI Mentor produces the final summary and next steps

Current UI

The current Streamlit app includes:

• A simple landing section
• Score cards for readiness, impact, language, and external PRs
• A mentor-first summary block near the top
• Profile metrics and score breakdown charts
• Repository cards with language, stars, forks, and update time
• A workflow panel showing the multi-agent pipeline

Tech Stack

• Backend: FastAPI
• Frontend: Streamlit
• Agent orchestration: LangGraph
• LLM: Groq via langchain-groq
• Data source: GitHub REST API

Setup

git clone <your-repo-url>
cd GitInsight
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Environment Variables

Create a .env file in the project root:

GROQ_API_KEY=your_groq_api_key_here
GITHUB_TOKEN=your_github_token_here

Notes:

• GROQ_API_KEY is required for the AI-generated code-quality and mentor outputs.
• GITHUB_TOKEN is recommended to avoid strict GitHub rate limits.

Run The App

Start the FastAPI backend:

uvicorn main:app --reload

Start the Streamlit frontend in another terminal:

streamlit run ui/app.py

Optional:

• Set GITINSIGHT_API_URL if your backend is running somewhere other than http://127.0.0.1:8000

Example:

export GITINSIGHT_API_URL=http://127.0.0.1:8000
streamlit run ui/app.py

API

GET /

Health check endpoint.

POST /review?username=<github_username>

Runs the GitHub analysis workflow and returns:

• profile
• repositories
• metrics
• details
• feedback
• workflow
• errors

Demo

Demo video:

demo_video/demo.mov