HvenBot — why it visited your site
If you saw Mozilla/5.0 (compatible; HvenBot/1.0; +https://hven.io/bot) in your server logs, this page explains exactly what fetched your page and why.
What HvenBot is
HvenBot is the page fetcher used by Hven, a tool that helps businesses see and improve how AI assistants talk about them. It fetches a page because of a person: a business setting up its Hven profile or adding one of its own pages for Hven to read, someone running a one-off visibility scan of a site, or — once a day — a check of whether a fix a business chose to make on its own site (such as publishing an /llms.txt file) is live yet. It never roams the web on its own.
What it actually does
- One page per request. It fetches the specific public page it needs — usually the homepage, or a file such as
/llms.txt— and stops. It does not follow links or crawl the rest of your site. - Public pages only. It reads what any browser sees. No logins, no forms, no cookies set on your visitors.
- Light touch. It reads a capped amount of each page, gives up after 30 seconds, and simply moves on if your server says no.
- No visitor data. It looks at page content, not at the people visiting your site.
- Not a search engine. Nothing it reads is published as a public index or archive.
HvenBot checks your robots.txt before every fetch and honors Disallow rules for HvenBot or *. If your robots.txt says no, the page is simply not fetched.
How to opt out
Disallow HvenBot in your robots.txt and every future fetch is skipped automatically:
User-agent: HvenBot
Disallow: /
Or email info@hven.io with your domain and a human will take care of it for you.
Seen Hven-PitchAudit instead?
That's one of Hven's other fetchers: a one-page fetch of a site's public homepage made when a marketing agency using Hven runs a one-off AI-visibility audit of that site. Same manners — one page, no link-following, no visitor data — and it honors robots.txt under Hven-PitchAudit or *. Disallowing Hven-PitchAudit blocks those audits without affecting anything a site's own owner asks Hven to read.
Seen Hven-OutreachDiscovery instead?
When Hven looks for marketing agencies it might contact, an AI assistant's answer sometimes cites a web page. This fetcher reads that one cited page, once, to see which businesses it names. It honors robots.txt under Hven-OutreachDiscovery or *.
Seen HvenOnboarding instead?
An older reader that fetches the homepage of a business that signed up to Hven, for its title and description, when Hven writes that business a welcome note. It reads one page and does not follow links. It does not check robots.txt; email info@hven.io if you want it to skip your site.
Seen HvenOutreachProbe instead?
That fetcher runs on Hven's own schedule, not because someone asked. It visits a business Hven is thinking of contacting, at most once a week, and looks only for an email address the business published itself as a mailto: link. It reads up to five pages: /, /contact, /contact-us, /about and /about-us. It reads your robots.txt first and skips any of those pages that the User-agent: * rules disallow. A person at Hven checks every address it finds before it is used.
Questions
Anything else — info@hven.io.