About Overfront
Overfront is a free Chrome extension that shows how visible a website is in AI-generated answers: an AI visibility score, a separate Brand score, the competing domains cited in the same answers, the discovery questions behind the result, and the sources those answers link to. It works on any public website, needs no account, and sends only the site's domain and an anonymous identifier. You can install it from the Chrome Web Store.
Who builds it
Overfront is built and written by U Hashimoto, an AI engineer with a background in shipping software products. U Hashimoto designs the measurement service, builds the extension, and writes the guides and articles published here. There is no separate content team, so the person explaining a result is the person who built the thing that produced it.
Why the measurement is independent
The service observes AI answers from the outside. It has no partnership with, and no data feed from, OpenAI, Google, Anthropic, or any other AI provider. Scores are our own estimates from sampled answers, not official rankings, and no AI provider endorses them. Our Method page explains what the two scores mean, what current measurement covers, and where its limits are. When the meaning, scope, or limits of a result change, that page changes in the same release.
We do not sell placements, do not write or automate reviews or forum posts on anyone's behalf, and do not accept payment to influence a score. A site that wants to be cited more often has to earn it on the page.
How the guides are written
Every guide and article on this site follows the same rules. Claims about how AI search works cite a primary source, such as a research paper, an engine's own documentation, or an observation we made ourselves and dated. Numbers we cannot verify are left out rather than estimated. Articles carry a named byline and a publication date, and when one is revised in a way that changes its advice, the update date is shown next to it. We publish a few pieces a month and revisit the core guides periodically rather than publishing in volume.
The starting point for new readers is the guide to generative engine optimization; the blog covers narrower questions.
What we commit to after launch
Launching an extension starts a relationship with the people who use it. We keep the experience close to the page you are already researching, keep the distinction between a discovery score and a Brand score clear, and explain waiting states rather than leaving you guessing. We describe only the engines and evidence that are actually supported today, and we say so when a change affects how a result should be read.
Contact
Questions about a result, a correction to an article, or a support request: [email protected]. Include the domain and the measurement date when asking about a score. Our privacy policy describes what the extension sends and what this website records.