AI search visibility for any domain. Enter a domain to see its Agent Citability score: a 0 to 100 measure of how deeply the trusted layer of the web references it.
Agent Citability scores a domain from 0 to 100 on how deeply it is woven into the trusted layer of the web: the reference works, government and academic sites, and editorial newsrooms that AI answer engines lean on when they answer a question. It is built to predict whether those engines would treat a domain as citable.
Links are weighted by who published them, from a curated list of over 1,000 sources. A reference from an encyclopedia or a university page moves a score. A listing anyone can add in a minute does not. Scores are calibrated within each release, and they can go down as well as up, which is the difference between this and the authority metrics it sits next to.
This is a v0 preview. The full method, the tier philosophy and what the score deliberately does not claim are on the methodology page, and the reasoning behind building it at all is in the announcement.
See the score and the trusted sources behind it
Agent Citability is one number and the evidence behind it. Here is what each part of a result means.
A 0 to 100 score for how deeply a domain is referenced by the parts of the web that answer engines treat as trustworthy. Higher means more of that reference, from better places.
Links from a curated, tiered list of over 1,000 sources: reference works, government and academic sites, wire services and editorial newsrooms. A link anyone can post in a minute counts for almost nothing.
The domains listed under a result are the trusted sources we found linking to it. They're the part you can go and verify yourself, which is why we show them rather than just the number.
Agent Citability measures citability, not citations. It does not predict that any engine will cite a domain, and it says nothing about whether the content on it is any good.
Current release: May–July 2026 index · preview
What answer engines cite, why that is measurable from links, and how to read the score.
AI answer engines draw on the sources that search engines and human editors already trust: reference works, government and academic sites, wire services, established newsrooms. Whether a domain is cited in any one answer depends on the query and the engine, and no tool can promise a citation. What can be measured is the layer underneath. Does the trusted part of the web already reference the domain? Agent Citability measures that.
We keep a curated, tiered list of over 1,000 source domains. For each domain you check, we look for links from those sources in the Common Crawl web graph and our own crawl, weight each link by the tier of the source that published it, and calibrate the total to a 0 to 100 scale within the release. A link from an encyclopedia or a university page moves a score. A directory listing anyone can add does not. The method, the tier philosophy and the release notes are published on the methodology page; the exact weights are not.
The score is a summary. The source list under it is the evidence. Every domain listed is a trusted source we found linking to the result, and you can open any of them and find the link. Read the sources before the number: three links from national newspapers and a university say more than a dozen from the lowest tier. The release label says which index the score came from, and a score only changes when a new release ships.
Unlike domain authority metrics, which mostly climb, this score can fall. It drops when a trusted source removes its link, when a source is re-tiered, or when a new release recalibrates the scale. A falling score is information. It means the trusted web references the domain less than it did.
We hold a score only for domains that at least one trusted source links to. If none does, there is nothing to measure and the page says so instead of showing zero. A zero would claim we measured and found nothing, which is a different statement.
An expired domain keeps the references it earned until the linking pages change. A lapsed name that a government site or a newspaper once cited still carries that reference, which is why citability is shown on directory listings alongside authority. Check the score here, then run the domain authority and backlink checkers on the same name to see the wider link profile behind it.
Agent Citability is a 0 to 100 domain score for how deeply a site is referenced by the trusted layer of the web: the reference, government, academic and editorial sources that AI answer engines draw on. It weights links by who published them rather than counting them all the same.
Enter the domain above. The score tells you how deeply the sources answer engines trust already reference it, and the list under the score names those sources. No tool can promise a citation, because that depends on the query and the engine, but a domain nothing trusted links to is rarely cited.
Yes. No signup, no card, and no usage cap beyond a per-visitor rate limit. Enter a domain and you get the score and the trusted sources behind it.
Domain Authority and Domain Rating estimate ranking strength from the shape of the whole link graph, and they mostly only go up. Agent Citability asks a narrower question: does the trusted layer of the web reference this domain? It can also fall, which is the point of it.
We only hold a score for domains that trusted sources link to. If nothing in that set links to a domain, there is nothing to score and we say so rather than showing a zero. A zero would claim we measured something and found it empty, which is a different statement.
Read next
Every listing shows its citability and authority alongside referring domains and archive history, before the name is revealed. Browse the directory of expired domains with backlinks.
