24 actively-maintained recommendation engine agents · ranked by maintenance health · 33,699+ indexed
Looking for the best open-source recommendation engine AI agent? Beacon indexes 33,699+ agents and MCP servers from GitHub and ranks them by real maintenance health — freshness, stars, and activity — so abandoned repos don't top the list. Save time: one search returns ranked, linkable agents in under a second (vs hours on GitHub). Below are 24 recommendation engine agents you can install today — free, no API key.
Use from your editor: BEACON_REGISTRY_URL=https://registry-ruby.vercel.app npx -y beacon-mcp — ask your AI to find a recommendation engine agent.
For autonomous agents (API)
Free, keyless discovery — paste into your agent workflow:
GET https://registry-ruby.vercel.app/api/v1/search?q=recommendation+engine&limit=10
POST https://registry-ruby.vercel.app/api/v1/discover
Body: {"query": "recommendation engine", "limit": 10}
What is the best open-source recommendation engine agent?
Top-ranked recommendation engine agents on Beacon include Headroom, Gpt Engineer, Awesome Ai Agents, ranked by maintenance health — recent activity, GitHub stars, and open-issue load — not stars alone.
Are these recommendation engine agents free to use?
Yes. Every recommendation engine agent listed is open-source and free. Beacon is a free directory — no API key or signup to browse or search.
How do I use a recommendation engine agent from my editor or AI assistant?
Run 'npx -y beacon-mcp' to add Beacon as an MCP server in Cursor, Claude Desktop, Cline, or Windsurf, then ask your AI to find a recommendation engine agent.
How does Beacon save time finding a recommendation engine agent?
Instead of searching GitHub and reading dozens of READMEs, call GET https://registry-ruby.vercel.app/api/v1/search?q=recommendation+engine — sub-second results ranked by health, stars, and activity.
Can autonomous AI agents discover recommendation engine tools programmatically?
Yes. Agents read https://registry-ruby.vercel.app/llms.txt and call POST /api/v1/discover with {"query": "recommendation engine"} — no API key required.