BetterYT

A Chrome extension that lets you create custom, curated video feeds directly on YouTube's homepage. Organize channels and videos into categories, and BetterYT surfaces a personalized feed with AI-powered recommendations — right alongside YouTube's native UI.

Features

• Custom category chips — Appear in YouTube's homepage chip bar. Click one to replace the native feed with your curated feed.
• Channel & video seeds — Add YouTube channels or individual videos as "seeds" to any category. BetterYT pulls recent uploads via RSS and discovers related content via the YouTube Data API.
• AI reranking — Optionally uses OpenAI (gpt-4o-mini) to score and filter recommendations for relevance to your interests.
• "Add to Feed" buttons — Injected on video pages, channel pages, and homepage thumbnails for quick one-click additions.
• Smart caching — Suggestion results are cached per category per day and automatically invalidated when seeds change.

Setup

Prerequisites

• Node.js (v18+)
• A YouTube Data API v3 key
• (Optional) An OpenAI API key for AI reranking

Install & build

npm install
npm run build

Load into Chrome

1. Go to chrome://extensions
2. Enable Developer mode
3. Click Load unpacked and select the dist/ folder

Configure API keys

1. Click the BetterYT extension icon to open the popup
2. Click the gear icon (⚙) to open Settings
3. Enter your YouTube Data API key (required) and OpenAI API key (optional)
4. Click Save Keys

Usage

1. Create a category in the extension popup (e.g. "Machine Learning", "Music Production").
2. Add channels or videos — browse YouTube and use the "Add to Feed" buttons on video/channel pages, or on homepage thumbnails. You can also add channels manually in the popup.
3. View your feed — go to the YouTube homepage and click your category chip in the filter bar.
4. Refresh suggestions — click the refresh button in the feed or in the popup to re-fetch recommendations.

Tech stack

• React 19 + TypeScript
• Vite + @crxjs/vite-plugin for Manifest V3 bundling
• Chrome Storage API for local persistence
• YouTube RSS feeds + YouTube Data API v3 for content
• OpenAI API for optional AI reranking

Project structure

src/
├── background/     # Service worker — RSS fetching, suggestion pipeline, caching
├── content/        # Content script — injects chips, feed UI, and "Add to Feed" buttons
├── icons/          # SVG icon components
├── popup/          # Extension popup — category management and settings
├── services/       # YouTube API client, recommendation engine, AI reranker
└── utils/          # Chrome storage helpers, types

Development

npm run dev      # Start Vite dev server with HMR
npm run build    # Type-check and production build to dist/
npm run lint     # Run ESLint