AI Changelog Generator

Generate detailed changelogs from GitHub pull requests using AI.

Features

• 🔍 Fetch PRs from any public GitHub repository
• 📅 Flexible date range selection (last N days or custom range)
• 🤖 AI-powered changelog generation with categorization
• 📝 Beautiful MDX rendering with syntax highlighting
• ⚡ Real-time streaming for instant feedback
• 📥 Download changelog as MDX file
• 🎨 Dark mode support

Setup

Environment Variables

For better GitHub API rate limits, add a GitHub token:

```bash
GITHUB_TOKEN=your_github_personal_access_token
```

Without a token, you're limited to 60 requests/hour. With a token, you get 5000 requests/hour.

Creating a GitHub Token

1. Go to GitHub Settings → Developer settings → Personal access tokens → Tokens (classic)
2. Generate new token with public_repo scope
3. Copy the token and add it to your environment variables

Usage

1. Enter a GitHub repository in owner/repository format (e.g., vercel/next.js)
2. Select a time frame:
   • Last N Days: Get PRs from the last X days
   • Date Range: Specify exact start and end dates
3. Click "Generate Changelog"
4. Wait for the AI to process and generate your changelog
5. Copy or download the generated changelog

Handling Large Repositories

Timeout Prevention

The app is designed to handle repositories with hundreds of PRs:

• API Route Timeout: Set to 60 seconds (maxDuration = 60)
• Fetch Timeout: 30 seconds total, 10 seconds per request
• PR Limit: Fetches up to 1000 PRs (10 pages × 100 PRs)
• Token Limit: Truncates PR descriptions to 500 characters
• Batching: For >200 PRs, only the first 200 are sent to AI with a note about truncation

Why Timeouts Might Occur

1. GitHub API Rate Limits: Without a token, you're limited to 60 requests/hour
2. Large Date Ranges: Querying years of PRs takes longer
3. Network Latency: Slow connections to GitHub API
4. AI Processing: Large amounts of data take longer to process

Solutions

1. Add a GitHub Token: Increases rate limits from 60 to 5000 requests/hour
2. Reduce Date Range: Query smaller time periods
3. Use Specific Dates: Target specific release periods instead of broad ranges
4. Deploy to Vercel: Better network connectivity and serverless function limits

Technical Details

Architecture

• Frontend: Next.js 16 with React Server Components
• Styling: Tailwind CSS v4 with custom design tokens
• AI: Vercel AI SDK v5 with streaming
• Markdown: react-markdown with GitHub Flavored Markdown support
• Syntax Highlighting: react-syntax-highlighter with Prism

API Endpoints

• POST /api/generate-changelog: Fetches PRs and generates changelog with streaming response

Rate Limits

The hosted version has a ratelimit of 5 req every 60seconds using Unkey. You can remove that and use your own Github token to increase the rate limit.

Fingerprinting for Rate Limiting

This application uses browser fingerprinting to identify unique users for rate limiting purposes:

• Technology: FingerprintJS is used to generate a unique browser fingerprint
• Purpose: Each unique browser fingerprint is rate limited to 5 requests per 60 seconds
• Privacy: The fingerprint is only used for rate limiting and is not stored permanently
• Implementation: The fingerprint is automatically generated when the form loads and sent with each request

The fingerprinting ensures fair usage across all users while maintaining privacy and not requiring user registration.