2c:I[75397,["/_next/static/chunks/0h0py-_snytkj.js","/_next/static/chunks/14d801gjinolu.js","/_next/static/chunks/0vv3t5x_ljlle.js","/_next/static/chunks/0u25v28zhnj_e.js","/_next/static/chunks/16c_778at-ygm.js","/_next/static/chunks/11ymd45ntieg-.js","/_next/static/chunks/07yr2l24mmpyf.js","/_next/static/chunks/01dh_eyj2m6ie.js","/_next/static/chunks/0d~6feex40k98.js","/_next/static/chunks/13~2t92r45w-x.js","/_next/static/chunks/0hhltp92k029x.js"],"UserView"] 2d:T506,More and more developer tools are adopting the llms.txt standard to build AI-friendly versions of their docs. The problem is that it's very hard to search across them. So I crawled millions of webpages to build Statespace, a search engine for llms.txt sites. And it's free, no API keys required.

You can run plain queries to search across all sites:

  mcp server setup
  vector database embeddings
  oauth2 refresh token

Or scope your queries to a specific site with "site: query"

  stripe: webhook verification
  mistral.ai: function calling
  docs.supabase.com: edge functions auth

Quotes work like Google for exact phrases:

  "context window limit"
  vector database "semantic search"
  stripe: "webhook signature verification"

This is still a work in progress, so looking for feedback and beta testers!

- Search for humans (website): https://statespace.com/

- Search for agents (CLI, SDK, MCP, skill): https://github.com/statespace-tech/statespace

投稿

Show HN: I built a search engine for llms.txt sites

[untitled]

コメント