VakaBot
VakaBot is the bot of Vaka · Website watch, a service that monitors websites for their owners. It only visits websites that someone added to Vaka, and only reads what they show publicly anyway.
How to recognise it
VakaBot always identifies itself with this user agent:
Mozilla/5.0 (compatible; VakaBot/1.0; +https://vaka.fyi/bot)
Requests come from Cloudflare's network. There are no fixed IP addresses for them; VakaBot is recognised by its user agent.
Why it visits your website
Someone added your website to Vaka, usually you or the agency that looks after it. VakaBot does not look for websites on its own and does not follow links on to other websites.
When a team also checks links to other websites, it only requests the linked address itself there, with a HEAD request.
What it requests and how often
- The address that was added, usually the start page: at most once a minute, depending on the team's settings.
- Once a day: the redirect from http to https, robots.txt and sitemap, the start page's scripts and stylesheets, a sample of up to ten links and, on WordPress, plugins' readme.txt and the theme's style.css.
- Addresses the team enters itself, such as an API like /api/health.
- Once a week, if the team turned it on: a walk through the website's pages (see below).
Performance is measured once a day by Google PageSpeed Insights. Those visits come from Google (user agent with “Chrome-Lighthouse”), not from VakaBot.
The weekly walk
- At night between 1 and 5 am in the team's time zone, at most as many pages as its plan allows (up to 10,000).
- Only one request at a time, at most one per second, or one every three seconds if the team prefers. If your server answers slowly, VakaBot waits correspondingly longer.
- It follows a Crawl-delay in robots.txt up to 10 seconds.
- On 429 or 503 it pauses, doubling the pause each time, and follows Retry-After (up to 10 minutes). After three such answers in a row it ends the walk.
- It reads HTML pages up to 2 MB. The files they load (images, scripts, stylesheets) it checks with HEAD requests only.
What it does not do
VakaBot submits no forms, tries no logins, does not probe for vulnerabilities, uses no cookies and runs no JavaScript.
Only results are stored: status, response times and findings with the page's address, plus a preview picture of the start page. We keep no copies of your pages.
Control it with robots.txt
The weekly walk follows robots.txt (RFC 9309): the rules for VakaBot, otherwise those for everyone (*). This keeps it out entirely:
User-agent: VakaBot Disallow: /
And this slows it down:
User-agent: VakaBot Crawl-delay: 5
People using Vaka can turn this off for their own website, for example to include areas closed to bots. If VakaBot ignores your robots.txt and you did not ask for that, write to us.
The regular checks of the added address do not read robots.txt, like any uptime monitor: someone set them up on purpose. If you don't want them, block the user agent “VakaBot” in your firewall or write to us.
Contact
A question, a complaint, or should your website no longer be checked? Write to us, we read every message.