The Topics API relies on a taxonomy of topics to categorize websites and user interests. This taxonomy is designed to be transparent, privacy-preserving, and relevant to the advertising ecosystem.
- Limited size: The taxonomy is intentionally small (currently 469 topics) to reduce the risk of fingerprinting and protect user privacy.
- Human-curated: The topics are not only human-readable, they are human-curated to ensure they are relevant to advertising and to avoid the inclusion of sensitive categories.
- Publicly visible: The list of topics is publicly shared so that users and the community can evaluate them and provide feedback.
- No demographic influence: User demographic information is not used to inform the classification model or the topics returned by the API.
How topics are assigned
- Only parties that have called the Topics API on a site matching a topic can receive that topic in the following weeks.
- All parties that interact with the Topics API on a given site will receive the same topic for a given week, if they are eligible to see it.
FAQs about the Topics API taxonomy
Q: How often will the taxonomy change?
A: Chrome will seek ecosystem feedback to shape the taxonomy. Longer term, Chrome would like the taxonomy to be provided by an external party that incorporates feedback and ideas from across the industry. See the public discussion on taxonomy for more information.
Q: Can API callers store Topics along with their first-party data?
A: In theory, yes. Callers can store or log the topics returned along with their first-party data. However, this is not an endorsed use case. Callers should not be using or relying on the topics returned by the API to try to re-identify the user across sites.