Read the signal
Homepage text tells us what a site does. New fetches focus on its title and description.
THE WEB, CONNECTED BY MEANING
A whole web of
similar possibilities.
Start with a website. Discover its neighbors.
Find competitors, explore a niche, or follow an idea
somewhere you haven’t been.
What’s your starting point?
Free to search. No account needed.Try a different direction
Discovering neighbors of stripe.com
LIVE INDEXLooking up this neighborhood…
| # | Website | Semantic similarity | DR |
|---|
SMALL FOOTPRINT. VERY BIG AMBITION.
One vector for every
English-language website.
A title. A description. A point in semantic space. We’re building an open index that connects websites by what they’re about, so anyone can discover the web beyond the usual names.
Millions of homepages are already searchable. The foundation comes from Common Crawl; direct homepage fetches are the next step toward broader, fresher coverage.
Explore how it’s builtFind the connection
Embeddings place related ideas near each other, even when the words are different.
Make it accessible
An open model, a compact index, and free public search. Use it here or build with it.
TAKE THE INDEX WITH YOU
Your tools.
The same open web.
Connect an MCP agent or make an HTTP request.
Public search returns up to 10 matches, without an API key.
Add this remote server in your MCP client.
One search tool: find_similar_sites