Give it a keyword, a domain, or a URL — get the most meaning-similar sites from across millions of homepages, each with a live authority score. Free, no signup, one line to install.
Drops into
Most similar in meaning
| # | Domain | Sim | DR |
|---|---|---|---|
| 1 | collectforstripe.com | 0.799 | 29.0 |
| 2 | sticky.io | 0.784 | 65.0 |
| 3 | chargehive.com | 0.782 | 4.5 |
| 4 | paymenthouse.com | 0.779 | 34.0 |
| 5 | payoneer.com | 0.771 | 88.0 |
Paste this to Claude (or Claude Code) — it connects the remote server itself, nothing to download or clone. Then ask it to find similar sites.
Add the Open Vector Index MCP server so you can find sites similar to any domain. It's a remote server at https://openvectorindex.com/mcp — transport "http", no API key needed. Once it's connected, use the find_similar_sites tool (try the query "stripe.com").
Prefer the terminal? Run claude mcp add --transport http openvectorindex https://openvectorindex.com/mcp
Paste this to Codex — or any coding agent that speaks MCP — and it wires itself up. No local file.
Add the Open Vector Index MCP server so you can find sites similar to any domain. It's a remote, streamable-HTTP MCP server at https://openvectorindex.com/mcp — no API key needed. Once it's connected, use the find_similar_sites tool (try the query "stripe.com").
Wiring it in by hand? Drop this into your MCP client config — one remote server, one tool, find_similar_sites. No local script to run.
{
"mcpServers": {
"openvectorindex": {
"type": "http",
"url": "https://openvectorindex.com/mcp"
}
}
}
Prefer the terminal? The ovi command is a single Python file — no SDK, no signup. Grab it and search from your shell.
curl -O https://raw.githubusercontent.com/joshstorz/vector-index/main/scripts/ovi chmod +x ovi OVI_API_BASE=https://openvectorindex.com ./ovi similar "vegan meal delivery"
No key needed for public search. Send a domain, a URL, or free text.
curl -s https://openvectorindex.com/search \
-H 'content-type: application/json' \
-d '{"domain":"stripe.com","k":10}'
Standard library only — no SDK to install.
import json, urllib.request
req = urllib.request.Request(
"https://openvectorindex.com/search",
data=json.dumps({"domain": "stripe.com", "k": 10}).encode(),
headers={"content-type": "application/json"})
print(json.load(urllib.request.urlopen(req)))
Runs anywhere fetch does — Node 18+, Deno, the browser.
const r = await fetch("https://openvectorindex.com/search", {
method: "POST",
headers: { "content-type": "application/json" },
body: JSON.stringify({ domain: "stripe.com", k: 10 }),
});
console.log(await r.json());
Type a keyword, a domain, or a URL — the same free search the CLI and agents use. No signup, no key.
Public endpoint · no key · rate-limited to keep it free · up to 10 results.
| # | Similar site | Similarity | DR |
|---|
One search, lots of jobs.
Drop in a company's domain and get its closest peers by what their homepage actually means — not by matching keywords.
"Sites like this one" surfaces topically-aligned sites worth a pitch, each with a live authority score.
Type a keyword or paste any URL and discover the neighbours in that corner of the web.
The semantic-match piece, in plain terms.
Exa is a full neural search engine — it answers questions and pulls passages from across the live web. Its homepage leads with an interactive playground and tucks integration into the docs.
Open Vector Index does one thing well: "what sites are most similar in meaning to this?" across millions of homepages. It's free, installs in one line, and works the same from the search box above, your terminal, or any AI assistant.
If you need open-web question-answering, reach for a search API. If you need fast, cheap, structured "sites like this" with authority scores attached — this is the smaller, sharper tool.