Topics API for Web overview

The Topics API enables interest-based advertising (IBA) without having to resort to tracking the sites a user visits.

Implementation status

The Topics API has completed the public discussion phase and is currently available to 99 percent of users, scaling up to 100 percent.
To provide your feedback on the Topics API, create an Issue on the Topics explainer or participate in discussions in the Improving Web Advertising Business Group. The explainer has a number of open questions that still require further definition.
The Privacy Sandbox timeline provides implementation timelines for the Topics API and other Privacy Sandbox proposals.
Topics API: latest updates details changes and enhancements to the Topics API and implementations.

What is the Topics API?

The Topics API is a Privacy Sandbox mechanism designed to preserve privacy while allowing a browser to share information with third parties about a user's interests. Learn how Topics enables enables interest-based advertising without having to resort to tracking the sites a user visits.

How it works

In the past, third-party cookies and other mechanisms have been used to track user browsing behavior across sites to infer topics of interest. These mechanisms are being phased out.

With the Topics API, the browser observes and records topics that appear to be of interest to the user, based on their browsing activity. This information is recorded on the user's device. The Topics API can then give API callers (such as ad tech platforms) access to a user's topics of interest, but without revealing additional information about the user's browsing activity.

Observation of ancestor topics

Since Chrome 114, when a caller observes a topic for a user on a page, the browser also regards the caller as having observed all ancestors of the topic.

For example, if the browser records that a caller observes Shopping/Apparel/Footwear/Boots for a user, then the ancestors of that topic are also regarded as having been observed: Shopping/Apparel/Footwear, Shopping/Apparel, and Shopping.

Previously, in order for a caller to be regarded by the browser as having observed (for example) Shopping/Apparel, that specific topic would have to have been returned by the API as the topic observed. This means that if Shopping/Apparel is observed for a caller for a user on one page, and Shopping/Apparel/Footwear/Boots on another, then the API will treat Shopping/Apparel as having been observed on both pages.

Epochs

Of course the Topics API must ensure that the topics of interest it provides are kept up to date. The browser infers topics for a user based on their browsing activity during a period of time known as an epoch, one week by default. Each user has their own epochs (epochs are "per user") and the initial start time is randomized. The topic selected for each epoch is randomly selected from the user's top five topics for that time period. To further enhance privacy and ensure that all topics may be represented, there is a 5% chance the topic is randomly selected from all possible topics in a taxonomy of interests.

The Topics API has three main tasks:

Map browser activity to topics of interest. With the current design of the Topics API, topics are inferred from the hostnames of pages the user visits. For example, the topic inferred for a website about aquariums might be /Pets & Animals/Pets.
Calculate the top topics for a user based on their recent browsing activity.
Provide mechanisms to access topics currently of interest to the user, to help select the appropriate ads.

The Topics API provides human-readable, easily understandable topics, so it's possible to provide meaningful controls to users.

How topics are curated and selected

Topics are selected from a taxonomy consisting of hierarchical categories such as /Arts & Entertainment/Music & Audio/Dance & Electronic Music and /Business & Industrial/Building Materials & Supplies. These topics have been curated by Chrome for initial testing, but with the goal that the taxonomy becomes a resource maintained by trusted ecosystem contributors. The taxonomy needs to be small enough that many users' browsers will be associated with each topic. We expect the final number of topics to be between a few hundred and a few thousand.

To avoid sensitive categories, topics must be public, human-curated, and remain up to date. The initial taxonomy proposed for testing by Chrome has been human-curated to exclude categories generally considered sensitive, such as ethnicity or sexual orientation.

For 50,000 top sites, the Topics API implementation in Chrome uses a manually curated, publicly available override list to map hostnames to topics. For other sites, the Topics API uses a machine learning model to infer topics from hostnames.

Chrome's implementation of the Topics API downloads a TensorFlow Lite file representing the model so it can be used locally on the user's device.

You can access the TensorFlow Lite model file and the topics inferred for hostnames from chrome://topics-internals.

The diagram that follows shows a simplified example to demonstrate how the Topics API might help an ad tech platform select an appropriate ad. The example assumes that the user's browser already has a model to map website hostnames to topics.

Diagram showing the stages in the Topics API lifecycle, from a user visiting websites to an ad being displayed. — The Topics API lifecycle diagram walks through the stages of the API actions from a high-level point of view.

API callers only receive topics they've observed

A design goal of the Topics API is to enable interest-based advertising without sharing information with more entities than is currently possible with third-party cookies. The Topics API is designed so topics can only be returned for API callers that have already observed them, within a limited timeframe. An API caller is said to have observed a topic for a user if it has called the document.browsingTopics() method in code included on a site that the Topics API has mapped to that topic.

The API returns only topics that have been observed by the caller within the most recent three epochs. This helps stop information about the user from being shared with more entities than the technologies the API is replacing (including third-party cookies).

The number of topics returned depends on the number of topics that the API caller has previously observed, and the number of topics that the user has available (such as the number of weeks of data accumulated). Anywhere from zero to three topics may be returned, as one topic can be indicated for each of the three recent epochs

For more information on how to use and test the Topics API, refer to the Topics API developer guide.

Next steps

Learn more about what topics are and how they work.
Read the developer guide for more in-depth resources.