Free educational Rankit tool

How Google categorizes your domain

Rankit Topic · Chrome Topics API reproduction (Google Privacy Sandbox)

For ad targeting, Chrome sorted every website into one or more topics, through the Topics API of Google's Privacy Sandbox. Rankit Topic faithfully reproduces both levels of that mechanism: for large sites it queries Chrome's official list (45,268 domains, exact result); for the rest, its engine estimates the topic from the domain name alone. Enter a domain, see the result.

What Rankit Topic reveals

The Topics API was part of the Privacy Sandbox, Google's project to replace third-party cookies for advertising. Chrome sorted every visited domain into a taxonomy of about 469 topics, to build the user's interest profile. Rankit Topic shows the result of that classification for any domain you choose.

What it is for

Beyond advertising, observing how a machine categorizes a domain from its name alone stays relevant for SEO and GEO: an inconsistent classification signals a domain name that is hard for engines to read, classic or generative.

Use cases

  • Understand a classification. See which sector the model files a domain into.
  • Compare a portfolio. Test several domains to spot inconsistencies.
  • Pick a name. Assess a domain's readability before buying it.
  • Catch a false positive. Identify a domain the model misclassifies.

Frequently asked questions

Does the Topics API still exist?

It is being retired. After abandoning the third-party cookie phase-out, Google announced the removal of the Topics API: deprecation with Chrome 144 (January 2026), full removal planned with Chrome 150. The document.browsingTopics() function will be removed. Rankit Topic keeps its value as an educational tool: it illustrates a classification mechanism still instructive for SEO and GEO.

Does the classification always come from a model?

No, and this is key. Google's documentation states topics are hand-curated for the most visited sites worldwide: for those domains, Chrome reads a mapping list, not the model. The machine-learning model only handles other domains. Rankit Topic faithfully reproduces both levels: it integrates Chrome's official list (45,268 domains, exact result, green badge) and, for any domain absent from that list, its engine estimates the topics (orange badge). Every result therefore clearly states whether it comes from the curated list or from an estimate.

Does Rankit Topic give exactly Chrome's result?

For the 45,268 domains in Chrome's official list, yes: Rankit Topic shows the exact result, integrated from that Google-curated list. For any other domain, Rankit Topic estimates the topics by rebuilding the model logic (hostname tokenization, voting against the taxonomy). The real model is a TensorFlow Lite file not reproducible without the file itself: the estimate is reliable, but not a bit-for-bit copy. Each result clearly states its source: exact or estimated.

What is the classification based on?

The domain name only, never page content, which is a privacy choice of the Topics API. Rankit Topic splits the name into tokens (for example boutiquevelo yields boutique and velo), then each known token votes for taxonomy topics.

Why do some domains get no topic?

An opaque or made-up domain name contains no recognizable word. Neither Chrome's model nor Rankit Topic can then reliably attach it to a topic.

Is my data sent anywhere?

No. Rankit Topic makes no network call: the classification runs entirely on the Rankit server, in a few milliseconds, without contacting Google or any third party.

Can I verify the result myself?

Yes, and it is even recommended. Chrome includes a built-in tool that shows the Topics API classification. On Chrome for desktop, copy chrome://topics-internals/ into the address bar, open the Classifier tab, then enter a hostname. For a domain in the official list, you will find exactly the result shown here. This tool only exists on Chrome desktop and will disappear when the Topics API is removed.

Why is my website not in the official list?

Chrome's official list is single and worldwide: it contains the most visited sites on a global scale, across all languages. Sites from any single country are therefore a minority, by simple proportion. A mid-sized website is very likely absent from it and handled by the estimation engine: this is not a flaw, it is the nature of this list. Only the very largest sites appear in the global list.

Sources

Analyze a domain

Enter a domain name or paste a full URL.

The www subdomain and any path are stripped automatically.
42 domains analyzed with Rankit Topic