NewThe detectors that scored perfect collapsed the hardest under attack.
Who we serve

The neutral party between the people who build detection and the people who depend on it.

Margen is independent of every detector on the market, so the same evidence carries weight on both sides. That neutrality is what lets one number serve three groups at once: the buyers deploying detection, the vendors building it, and the partners we publish and co-deliver with.

How the market fits together

One independent layer, trusted on both sides.

Vendors build detection. Buyers depend on it. We are the neutral party in the middle that measures whether it holds, with no detector of our own to sell.

They build it

Detection vendors

The models sold as fraud defense.

The neutral party

The measurement layer. We test whether the detection holds, with no detector of our own to sell.

They depend on it

Buyers

KYC, hiring, and enterprise security teams.

Partners stand alongside: research and red-team firms co-publish the benchmarks and co-deliver evaluations with us.
14
detectors evaluated
12
demographic groups covered per detector
1.00 to 0.34
top detectors, from perfect to below a coin flip
0
detectors we sell, by design

Every side of the same arms race comes to us for the same thing: a measurement of whether the detection holds. Find the one that fits you.

Buyers

You depend on detection you bought or built.

You hold real biometric data, and a detection system is the control standing between it and fraud. You need an independent number you can defend to auditors, procurement, and your own board.

Detector vendors

You build detection. Show that it holds.

Your buyers no longer take your own benchmark at face value. We supply third-party evidence, grounded in a corpus and a method your team did not control, that you can hand straight to procurement.

Red team

You extend your engagement into the detection layer.

Security and social-engineering firms co-deliver our detection-layer evaluation inside their own engagements. The customer gets one joint report covering both the human layer you already test and the technical layer we measure.

Academic

You help build the benchmarks everyone can cite.

Academic and standards organizations work with us on the open benchmarks and the shared measurement layer. The methodology stays public, so the results carry weight no single vendor could claim alone.

What we are

Margen is the measurement layer between detection vendors and the buyers who depend on them.

An independent assessment layer.

Third-party, with no detection product of our own to sell.

A reproducible methodology.

Every claim is backed by a dataset, a pipeline, and a margin of error.

Adversarially honest.

We test under the hardest conditions buyers face in production.

See where your detection system stands.

Whichever side of the market you are on, the question is the same: does your detection hold against what is actually coming? Bring it to us and find out, with the evidence behind the answer.