ZevianBot, explained.

ZevianBot is operated by Kishan Ashokbhai Kushwaha as part of Zevian, a product that audits a website, opens pull requests that fix SEO issues, and measures the result with the owner's own Search Console and Bing data.

It visits a site when someone has added that domain to a Zevian workspace. Anyone with a Zevian account can add a domain for an audit, so a visit does not by itself mean the site's owner asked for it. The audit is read-only: ZevianBot only requests pages, and it obeys yourrobots.txt as described below. If you did not want it, block it or tell us, and we will remove the domain.

Every request carries this user agent, which links to this page:

Requests are plain GET and HEAD requests. ZevianBot does not run JavaScript, send cookies, log in, or submit forms. It reads the HTML your server returns, the way a search engine's first pass does.

• Your home page, and from there up to 50 pages on the same site that it finds through links and sitemaps.
• /robots.txt, /llms.txt, and your sitemaps: those robots.txt declares, plus /sitemap.xml and /sitemap_index.xml.
• Up to 200 more internal URLs that it only checks, without keeping the page, to see whether they return errors or redirect. It tries HEAD first and uses GET only to confirm a bad status.
• At most 4 requests at a time, each given 10 seconds, following at most 5 redirects.

It audits each added site about once a day. For a few hours after a pull request that Zevian opened has been merged, it checks again as often as every 30 minutes to see whether the fix is live. The person who added the site can also start an audit by hand.

ZevianBot follows the robots.txt group that names ZevianBot. If there is no such group, it follows the * group. If there is neither, nothing is restricted.

Rules written for other bots, such as Googlebot or GPTBot, never decide what ZevianBot does. Zevian reports them to the site's owner as findings about what those bots can see.

It reads robots.txt before it requests anything else from a host, including a host that a redirect leads to, and it never requests a URL it is not allowed to. /robots.txt itself is always fetched. It does not apply Crawl-delay; it keeps to the request limits above.

When robots.txt cannot be read, ZevianBot follows RFC 9309. If the file is missing (the server answers with a 4xx, such as 404), nothing is restricted. If the server returns an error (5xx) or does not answer, ZevianBot crawls nothing on that site: it assumes the whole site is off limits until the file can be read, and asks for /robots.txt only. Zevian then records a critical finding for the site's owner, because search engines may treat the whole site as disallowed in the same situation. It tells the owner, once, that the problem is on the site's side.

Zevian asks again after 30 minutes, then after an hour, and then stops: at most 3 attempts in 24 hours, so a failing robots.txt is not hammered. The next scheduled audit starts a new round.

To keep ZevianBot out of the whole site, add this to your robots.txt:

To keep it out of part of the site, disallow just that path, for example Disallow: /private/. To let it in while other bots are blocked, give it its own group with Allow: /: a group that names ZevianBot replaces the * group for it.

It takes effect the next time ZevianBot visits; each added site is audited about once a day. A site that disallows ZevianBot is not crawled at all: the audit stops after reading robots.txt and tells the person who added the domain which rule said no.

You can also write to support@zevian.tech with the domain. We remove it from Zevian and stop visiting it, and reply within two business days. If it is causing load right now, say so in the subject line. If you added the domain yourself, you can delete the project in your Zevian workspace.

Bot protection, such as Cloudflare's Bot Fight Mode, can challenge ZevianBot. The audit then reports that the site refused the crawler instead of scoring an error page. To allow it, allow requests whose user agent contains ZevianBot.

We do not publish IP address ranges.

For the person who added the site: the audit results (issues and scores) and a summary of each page it read, meaning the title, meta description, canonical URL, H1, and a text excerpt of up to 400 characters. ZevianBot does not collect pages to build a search index or to train models. See the Privacy Policy for how that data is handled.