PR & Issue Summarizer Dashboard

A web application that fetches GitHub PRs/Issues and uses Claude AI to generate summaries, risk tags, and test checklists.

Tech Stack

Frontend: Next.js 14 (TypeScript) with Tailwind CSS
Backend: FastAPI (Python)
APIs: GitHub REST API, Claude AI via Amazon Bedrock
Auth: GitHub Personal Access Token (optional for public repos)
Deployment: Vercel (frontend) + AWS App Runner (backend)

Project Structure

pr-summarizer/
├── frontend/                    # Next.js application
│   ├── src/
│   │   ├── app/                 # App router pages
│   │   ├── components/          # React components
│   │   ├── lib/                 # API client
│   │   └── types/               # TypeScript types
│   ├── vercel.json              # Vercel deployment config
│   └── package.json
│
├── backend/                     # FastAPI application
│   ├── app/
│   │   ├── main.py              # FastAPI entry point
│   │   ├── config.py            # Environment config
│   │   ├── routers/             # API endpoints
│   │   ├── services/            # Business logic
│   │   └── models/              # Pydantic schemas
│   ├── requirements.txt
│   └── Dockerfile
│
├── scripts/                     # Deployment scripts
│   └── deploy-backend.sh        # ECR/App Runner deployment
│
├── docs/                        # Documentation
│   └── AWS_DEPLOYMENT.md        # Detailed AWS deployment guide
│
└── docker-compose.yml           # Local development

Prerequisites

Python 3.11+
Node.js 18+
AWS credentials with Bedrock access (for Claude AI)
GitHub Personal Access Token (optional — only needed for private repos)

Local Development

Backend

cd backend

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Create .env file
cp .env.example .env
# Edit .env and set AWS_REGION (requires AWS credentials with Bedrock access)

# Run development server
uvicorn app.main:app --reload

The backend will be available at http://localhost:8000.

Frontend

cd frontend

# Install dependencies
npm install

# Create .env.local file
cp .env.local.example .env.local

# Run development server
npm run dev

The frontend will be available at http://localhost:3000.

Using Docker

# Run backend with Docker Compose
docker-compose up --build

API Endpoints

Health Check

GET /health

Pull Requests

POST /api/v1/repos/{owner}/{repo}/pulls           # batch response
POST /api/v1/repos/{owner}/{repo}/pulls/stream     # SSE streaming
POST /api/v1/repos/{owner}/{repo}/pulls/{pr_number}

Issues

POST /api/v1/repos/{owner}/{repo}/issues           # batch response
POST /api/v1/repos/{owner}/{repo}/issues/stream    # SSE streaming
POST /api/v1/repos/{owner}/{repo}/issues/{issue_number}

All endpoints accept a JSON body with optional github_token field (required for private repos).

API Usage Examples

# Health check
curl http://localhost:8000/health

# Fetch PRs with summaries
curl -X POST http://localhost:8000/api/v1/repos/facebook/react/pulls \
  -H "Content-Type: application/json" \
  -d '{"github_token": ""}'

# Fetch PRs with streaming (SSE)
curl -N -X POST http://localhost:8000/api/v1/repos/facebook/react/pulls/stream \
  -H "Content-Type: application/json" \
  -d '{"github_token": ""}'

# Single PR summary
curl -X POST http://localhost:8000/api/v1/repos/facebook/react/pulls/123 \
  -H "Content-Type: application/json" \
  -d '{"github_token": ""}'

Deployment

Backend (AWS App Runner)

Automated via GitHub Actions on push to main (when backend/ files change).

Initial setup (one-time):

Run ./scripts/deploy-backend.sh --create-app-runner to create ECR repo and App Runner service
In App Runner console, set environment variables: ALLOWED_ORIGINS, AWS_REGION
Note the App Runner service URL and ARN

Required GitHub repo secrets:

Secret	Description
`AWS_ACCESS_KEY_ID`	IAM user access key
`AWS_SECRET_ACCESS_KEY`	IAM user secret key
`AWS_REGION`	AWS region (default: us-east-1)
`APP_RUNNER_SERVICE_ARN`	ARN from App Runner console

After setup, every push to main that changes backend/ files auto-deploys.

Frontend (Vercel)

Import the repo in Vercel
Set root directory to frontend
Add environment variable: NEXT_PUBLIC_API_URL = your App Runner URL
Deploy — Vercel auto-deploys on every push to main

Manual Backend Deployment

./scripts/deploy-backend.sh              # push to ECR only
./scripts/deploy-backend.sh --create-app-runner  # also create App Runner service

For detailed AWS setup instructions, see docs/AWS_DEPLOYMENT.md.

Environment Variables

Backend (.env)

AWS_REGION=us-east-1
ALLOWED_ORIGINS=http://localhost:3000,https://your-vercel-domain.vercel.app
LOG_LEVEL=INFO

Frontend (.env.local)

NEXT_PUBLIC_API_URL=http://localhost:8000

Features

PR Summaries: Concise summaries of pull request changes
Risk Tags: Automatic detection of breaking changes, security issues, performance impacts, etc.
Test Checklists: AI-generated test suggestions for each PR
Issue Analysis: Priority assessment and action items for issues
SSE Streaming: Results appear incrementally as each summary completes
No Token Required: Works with public repos without a GitHub token
Copy to Clipboard: Easy sharing of summaries
Recent Repos: Quick access to previously viewed repositories

Security Notes

GitHub tokens are stored in browser localStorage (can be cleared via UI)
Tokens are only sent to the backend for API calls, not stored server-side
Backend requires CORS configuration for production domains

GeezerVison

Latest commit

History

Folders and files

Repository files navigation