Get the freshest picks delivered straight to your Telegram — free, no sign-up, like a newsletter.
👉 Join the channel: t.me/ai_repos_daily
One tap to join, then every day you automatically receive a compact digest:
| What you get | Source |
|---|---|
| ⭐ Top GitHub repos of the day | Daily Top Picks |
| 🚀 Product Hunt winner of the day | Daily Top Picks |
| 🧠 Top paper + trending models | Hugging Face Daily |
| 📈 Breakout picks from the market report | AI Opportunity Report |
| 💡 Fresh startup ideas | Find Ideas |
| ⚙️ Automation radar stats | Automation Radar |
No spam, no noise — just one daily summary of what's heating up in AI.
You star an AI repo on GitHub. Then another. Then 500. Now your starred list is a black hole — no search, no filtering by category, no way to see what's trending.
AI Repositories turns your starred list into a fully searchable, filterable, trend‑aware directory — with AI analysis built in.
| Feature | What it does |
|---|---|
| 1,000+ AI repos | Hand‑curated, auto‑updated daily across 10 categories |
| 10 categories | Agents, Models & Inference, Developer Tools, Vision & Media, Infrastructure, AI Engineering, Data & Evaluation, Research & Knowledge, Applications, and more |
| Growth signals | 🔥 (500+ stars/week), ⚡ (100+/week), 📈 (positive momentum) — see what's heating up right now |
| Sparkline charts | 30‑day star history visualised for every repo |
| Stack Layer filter | Filter by infra → framework → tool → product → data → creative |
| Weekly Top Movers | 📊 Full ranking page: sort by delta, %, trend score + Top Gainers card view |
| Similar repos | In the drawer — finds repos with matching category, topics, language, star range |
| Idea Matrix | 10 industries × 10 AI features grid — click to generate startup ideas via LLM7 |
| Ask AI | Select repos → one‑click prompt for Google AI Mode → instant analysis |
| 12 prompt templates | Discovery, Build & Ship, Market Intelligence, Tech Stack + custom prompts |
| Find Ideas | Aggregates startup ideas from Show HN, Product Hunt, IndieHackers, Reddit, App Store, YouTube, V2EX, 36Kr |
| Live Market Report | Click "Generate" → LLM7 analyses all 1,000+ repos → structured market report with opportunities, trends, gaps, risks |
| Research Radar | Extended tech landscape with deeper analysis |
| Responsive | Desktop + mobile, dark theme, no sign‑up |
This isn't just a directory. It's a market intelligence tool:
- What's trending now? → Weekly Top Movers page with real growth data
- Where's the whitespace? → Generate a Live Market Report that identifies market gaps
- What should I build? → Idea Matrix + Find Ideas page with validated startup concepts
- Which repos are similar? → Built-in similarity engine in the detail drawer
A GitHub Actions pipeline runs every day:
- Fetches fresh stats for 1,000+ repos
- Computes growth deltas, trend scores, signals
- Pulls startup ideas from 10 sources
- Generates the AI Opportunity Report
- Deploys to GitHub Pages — no manual work ever
scripts/run_local_update.py crawls GitHub, classifies repos with a local LLM
(LM Studio on port 1234 or Ollama on 11434), rebuilds the agent-skills and
video-production datasets, then commits & pushes.
- Resume-aware: every repo is checkpointed (
data/history/local_update_state.json); if the machine sleeps/shuts down mid-run, re-running continues where it stopped. - LM Studio down: the wrapper waits up to 45 min for the endpoint, then exits;
re-run later resumes. The configured
LLM_MODELfalls back to any loaded model. - Auto-run (login + every 30 min, launchd coalesces missed runs after wake):
cp scripts/com.hankbui.starred-ai-update.plist ~/Library/LaunchAgents/
launchctl load ~/Library/LaunchAgents/com.hankbui.starred-ai-update.plistManual run: python3 scripts/run_local_update.py · force now: --now · custom wait: --wait 120
This site is built to be found:
| Tactic | Implementation |
|---|---|
| Semantic HTML | Proper <header>, <main>, <section>, <footer>, <h1>–<h3> hierarchy |
| Structured data | Each repo card uses semantic classes for potential schema.org indexing |
| Fast load | Pure static site — zero JS framework, zero build step, instant first paint |
| Mobile‑first responsive | Core Web Vitals‑friendly, responsive on all devices |
| Internal linking | Every page links to all others (index, weekly rank, ideas, report, radar) |
| Open source | Public repo indexed by Google — README, issues, discussions all crawlable |
| Auto‑updated | Fresh content daily = crawl frequency signal |
| Unique content | Original growth data, AI‑generated market reports, curated curation — not a copy |
Keywords this ranks for: Best AI repositories, Open source AI directory, GitHub AI projects 2026, AI tools list, Machine learning repos, Trending AI GitHub, LLM open source directory, AI startup ideas 2026, AI agent repositories, Weekly top AI GitHub repos
Vanilla JS · HTML · CSS — zero frameworks, zero dependencies.
Data served as static JSON. Hosted on GitHub Pages.
AI analysis via LLM7 API (free, no key required).
Want to help this project grow? Here's how (and why):
- ⭐ Star the repo — it's the single biggest signal for GitHub discovery
- 🔗 Link to it from your blog, newsletter, or "awesome‑AI" list
- 🐦 Share on X/Twitter — tag @hankbui and I'll reshare
- 🎯 Post on HN/Reddit — "I got tired of losing AI repos in my starred list, so I built this" resonates
- 📝 Add a missing repo — open an issue with
repo: owner/nameand it'll be included in the next daily update - 🤖 Use the Ask AI feature — every share or mention drives organic discovery
Viral loop: User discovers repos → uses Ask AI → gets value → shares → new users → repeat.
No code needed. To add a repo, open an issue with:
repo: owner/repo-name
category: (optional — we'll classify it)
Or submit a PR that updates the data source. All contributions welcome.
MIT — use the code, fork the site, build your own directory.
Built with ❤️ by @hankbui · Auto-updated daily · Free forever



