The Inovvo Site Classifier

Updated October 2026

If you found InovvoSiteClassifier in your server logs, this page explains who we are, why we visited, what we requested and how to opt out.

What it does

The Inovvo Site Classifier works out what kind of website a host name belongs to (for example a news publisher, an online shop, a streaming service or a software provider). The answers go into the Inovvo categorization database behind our Categorizer, which network operators use to understand and report on the traffic on their networks.

Being in the database matters for your site. A site we have categorized is recognized correctly in our customers' reports and services. A site we could not visit is listed as unknown, or is categorized from less reliable signals.

A minimal approach

We only need to know what kind of site you run, not what is on every page, so we ask for as little as possible:

• Your home page, and usually nothing else. Most sites receive one or two requests in total.
• robots.txt first. We read it once before anything else and follow it.
• Rarely. Once we have checked a site we reuse the result and do not come back for at least a week.
• Gently. Requests are spread out over time. We never send bursts of requests to one site.
• Read only. We never log in, fill in or submit forms, click advertisements, buy anything or solve security checks. If a site asks us to prove we are human, we stop.

The three ways we may visit

Every site gets the first step. Only sites the first step cannot decide get the second, and only a small sample gets the third.

1. Availability check

We look up your host name in DNS, read robots.txt and request your home page once over HTTPS (or HTTP if that is all your site offers), following redirects. From the response we keep only technical facts: the status code, the size and type of the response and its basic structure. We do not keep the page text.

2. Page reading

If the availability check cannot tell what kind of site it is, we load your home page once to read its public text, as a visitor would see it.

3. Visual check

To check the quality of our results, a small sample of sites is opened in a web browser, which takes screenshots of the home page. If the home page alone does not show what the site does, the browser may open up to three more pages on the same site, such as an "About" page, and only where robots.txt allows. Cookie banners are answered by rejecting non-essential cookies, never by accepting them.

How to recognize us

All three activities identify themselves. The availability check and page reading send this user agent:

Mozilla/5.0 (compatible; InovvoSiteClassifier/1.0; +https://inovvo.com/categorizer/classifier)

The visual check uses a real browser, so it sends that browser's normal user agent with our name added at the end:

... Chrome/... Safari/537.36 InovvoSiteClassifier/1.0 (+https://inovvo.com/categorizer/classifier)

Every request also carries the header From: info@inovvo.com. If you would like to verify that a request came from us, email us the time and the address it came from.

How to opt out

Add these lines to your site's robots.txt:

User-agent: InovvoSiteClassifier
Disallow: /

We follow the robots.txt standard (RFC 9309): we obey the group that names InovvoSiteClassifier, or the * group if none does. You can also disallow only some paths. If robots.txt disallows your home page, we do not request it, and we do not ask again until our next check of your robots.txt. Your site then stays out of content-based categorization, so it may be listed as unknown.

What we keep

• The category we assign to your host name.
• Technical facts from the availability check: the status code, the response size and type, and whether robots.txt allows us.
• For page reading, a short excerpt of the public text of your home page, used to decide the category.
• For the visual check, the screenshots, kept internally for quality review.

We do not collect personal data, we do not accept cookies beyond what a page requires, and we do not sell or publish your pages.

Contact us

If we have categorized your site wrongly, if our requests cause you any trouble or if you have a question, email info@inovvo.com. Please include your host name. We are happy to correct a category.