The reader behind Celorus research.
CelorusBot opens a public web page only when someone using Celorus asks for research on a company or a person, and only pages a search service returned for that request.
How to spot it
- It names itself in every request:
CelorusBot/1.0 (+https://celorus.com/bot). - It runs on Amazon Web Services and has no fixed set of addresses.
- One research request opens at most 12 pages, one after another.
- It follows no links from your page, and at most 3 redirects for a page.
- It reads web pages and plain text, nothing else.
Your robots.txt
Before each page, it reads your robots.txt and notes what the rules for CelorusBot, or the rules for every bot where none name it, say about that page. Today it opens the page either way. To keep it off your site, write to us (below).
What it keeps
It keeps no copy of your page. An AI model reads the pages and writes one summary for the person who asked. Where the summary uses your page, it names the page's address and the date it was read. That summary is a page on that person's own Celorus desk. CelorusBot's own log holds no address and no page text, and is kept for 30 days.
How to stop it
Write to tech@celorus.com with your site's name, such as example.com, and a person answers. We add it, and every name under it, to the list of sites CelorusBot never opens. On a shared host such as github.io, we add only your own site, such as yourname.github.io.