960+ hand-curated, production-tested GitHub repositories curated by Akash Priyadarshi.
Optimized for human developers exploring top tools and autonomous AI coding agents (Claude Code, Codex, Cursor, OpenClaw) needing instant, token-efficient project context.
Explore Web Dashboard • Read USAGE.md • API Reference • Categories Manifest
| 👤 For Humans | 🤖 For AI Agents & Vibe Coders |
|---|---|
| Compact List & Card Modes — Dense single-line table view vs modern card grid | Vibe Coder Project Matcher — Zero-latency client-side relevance engine matching your README/idea to top repos |
Local Bookmarks & Pinning — Save your shortlist to localStorage with zero login |
REST Search API — Serverless /api/search endpoint with CORS for instant agent queries |
1-Click git clone Copy — Instant clone commands copied directly to clipboard |
Per-Category llms-*.txt — Dedicated token-optimized endpoints for individual domains |
Instant Substring Search — Fuzzy search with / keyboard shortcut and live filter |
MCP Compatibility — Drop-in tool definitions for Model Context Protocol agents |
| Topic Clusters Matrix — Clickable tags (AI Agents, MCP, Claude Skills, Rust CLI, Android) | Zero-Cost Architecture — Vanilla JS + Edge serverless, zero database, zero external dependencies |
- Bloom warm-amber "aura" theme: atmospheric dark UI with drifting amber/coral blooms, film grain, giant weighty-sans display type (Inter Tight) + Instrument Serif accent, and refined shadcn-grade controls.
- Project Relevance Matcher: Paste your project idea or README. The client-side engine calculates TF-IDF and weighted keyword overlap scores, ranking the top matching tools with percentage match pills (e.g.
98% Match). - Keyboard-First Navigation: Press
/anywhere to focus search; pressEscapeto reset. - Progressive Batch Rendering:
IntersectionObserverrenders 60 cards per chunk, reducing initial DOM nodes from 7,100+ to ~450 for silky smooth 60fps scrolling and optimal Core Web Vitals (INP/LCP). - SEO & GEO Optimizations: Valid Schema.org JSON-LD graph (
WebApplication+Personentity graph), pre-rendered FAQ citation block for Google AI Overviews & Perplexity, and XML sitemaps.
Agents, scripts, and CLI tools can query the serverless endpoint directly:
GET https://my-starred-repos.vercel.app/api/search?q={query}&category={category}&lang={lang}&limit={limit}# Query AI agent and MCP tools
curl -s "https://my-starred-repos.vercel.app/api/search?q=mcp&limit=5"
# Query high-performance Rust tools ( slugs like dev-tools / ai auto-normalize to display names )
curl -s "https://my-starred-repos.vercel.app/api/search?category=dev-tools&lang=rust&limit=10"{
"schema_version": "0.2",
"total_tracked": 964,
"total_matches": 15,
"limit": 5,
"query": "mcp",
"category": null,
"language": null,
"results": [
{
"full_name": "modelcontextprotocol/servers",
"stars": 45120,
"language": "TypeScript",
"description": "Model Context Protocol Servers",
"category": "AI & Agents",
"url": "https://github.com/modelcontextprotocol/servers",
"last_updated": "2026-08-27T10:14:02Z",
"license": "MIT",
"forks": 4120,
"topics": ["ai", "mcp", "agent", "protocol"]
}
]
}Add this tool definition to Claude Code, Cursor, or your MCP agent configuration:
{
"name": "search_starred_repos",
"description": "Search 960+ hand-curated high-performance GitHub repositories across AI agents, MCP, Rust CLI, Android internals, and developer tools.",
"parameters": {
"type": "object",
"properties": {
"query": {
"type": "string",
"description": "Search query (e.g. 'mcp server', 'rust grep', 'android binder')"
},
"category": {
"type": "string",
"description": "Optional category: 'AI & Agents', 'Dev Tools', 'Mobile', 'Security', 'Web Automation', 'LLM & RAG', 'Web Dev', 'Other', 'Media', 'Databases', 'DevOps' (slugs like 'ai', 'dev-tools' also accepted)"
},
"limit": {
"type": "number",
"description": "Max results (1-50, default 10)"
}
},
"required": ["query"]
}
}| Endpoint | Description |
|---|---|
/llms.txt |
Top 100 Starred Repos + Topic Taxonomy + Category Manifest |
/llms-full.txt |
All 960+ Starred Repos with full metadata |
/llms-ai-agents.txt |
AI & Agents Category (539 repos) |
/llms-dev-tools.txt |
Dev Tools & CLI Utilities (74 repos) |
/llms-mobile.txt |
Android & Mobile Internals (50 repos) |
/llms-web-automation.txt |
Web Automation & Scraping (29 repos) |
/llms-llm-rag.txt |
LLM & RAG Frameworks (16 repos) |
/llms-web-dev.txt |
Web Dev Frameworks (11 repos) |
/llms-security.txt |
Security & Pentesting (10 repos) |
/llms-other.txt |
Other / Unclassified (207 repos) |
/llms-media.txt |
Media & Video/Audio (21 repos) |
/llms-databases.txt |
Databases & ORMs (2 repos) |
/llms-devops.txt |
DevOps & Infra (5 repos) |
| Category | Repos | Primary Technologies | Description |
|---|---|---|---|
| AI & Agents | 530+ | Python, TypeScript, Rust | Multi-agent frameworks, Claude Code skills, MCP servers, autonomous agents |
| Dev Tools | 70+ | Rust, Zig, Go, C++ | High-performance CLI tools, terminal emulators, compilers, token-efficient tools |
| Mobile | 50+ | Kotlin, Dart, Swift | Android internals, TDLib native builds, Flutter offline-first architecture |
| Web Automation | 30+ | TypeScript, Python | Headless browsers, crawler APIs, anti-bot scrapers, Playwright utilities |
| LLM & RAG | 16+ | Python, TypeScript | Vector embeddings, RAG pipelines, semantic search engines |
| Web Dev | 11+ | Next.js, React, Tailwind | Full-stack frameworks, design systems, modern web tools |
| Security | 10+ | Rust, Python, Go | Reverse engineering, forensics, memory inspection, vulnerability testing |
| DevOps & DBs | 10+ | Go, Shell, SQL | Docker, Kubernetes, SQLite/Postgres ORMs, infrastructure as code |
GitHub API (User Stars)
│
▼
┌─────────────────┐
│ fetch_stars.py │ ◄─── Auto-syncs weekly (Sunday 00:00 UTC) + on-demand
└────────┬────────┘
│
▼ writes repos_output.json (Single Source of Truth)
┌─────────────────┐
│ build_site.py │ ───► index.html (Baked SEO + Project Matcher UI)
└────────┬────────┘ ───► llms.txt, llms-full.txt, llms-*.txt
│ ───► sitemap.xml
▼
Vercel Edge CDN + Serverless API (/api/search)
# 1. Verify script syntax
node --check app.js api/search.js
python -m py_compile fetch_stars.py build_site.py generate_assets.py
# 2. Rebuild site artifacts
python build_site.py
# 3. Start local development server
python3 -m http.server 8080
# Open http://localhost:8080CUSTOM • Built with pride by Akash Priyadarshi.