A probability score, not a verdict.
Klyver checks the text a candidate submits and tells you how likely it is AI-generated, plus a fit score against the role. The signal sits on the candidate's card; you decide what to do with it.
Cover letters, answers to screening questions, written exercises, long candidate messages.
It isn't a filter that rejects on its own. It's one more signal, next to the CV, the experience and the interviews.
You know which texts are worth a closer look and which aren't.
You catch template answers before booking a call.
The team looks at the same signal, not the hunch of whoever reviewed.
From the portal, a form, or an answer to a screening question.
Flags the doubtful stretches, computes the AI score and the fit score against the role.
You read the signal alongside the rest of the profile and decide whether they advance, you ask more, or you pass.
Almost no ATS has this, and standalone detectors are often wrong about people writing in a second language. So in Klyver it's a data point inside the process, not a button that rejects. The call stays yours, with more to go on.
From analysing the text the candidate submits. It doesn't query external databases or send the content anywhere outside Klyver.
No. It never moves or drops anyone on its own. It only leaves the signal on the card.
Yes, both.
Yes, which is why we treat it as a signal, not a verdict. Read it with the rest of the profile.
From Growth for agencies and from Business for corporate teams.