We build the models that detect synthetic media — and we give them away free.

Detect Deepfakes is a free tool from Resemble AI, the generative-voice company that also happens to build one of the most accurate deepfake detectors in the world. If we're going to ship the technology that makes voice cloning effortless, we'd better be the ones giving everyone the tools to spot abuse.

Who uses this site

Journalists verifying a viral clip before publishing. Trust & safety teams triaging user-reported content. Fraud analysts confirming a suspicious voicemail. Parents checking a “call from the grandkid” before wiring money. If you need a second opinion on whether a piece of media is real, you're in the right place.

How detection works

Our models look for the statistical fingerprints left behind by synthesis pipelines — spectral anomalies in cloned voices, frequency-domain artifacts in diffusion-generated images, temporal inconsistencies in face swaps. No single tell is conclusive, so we fuse dozens of signals into a probability score.

The same models power detection for Netflix, Paramount, Deutsche Telekom, and a roster of banks and governments that we can't name publicly. On this site they're free. Through the Resemble AI API they run at production scale with SOC 2, GDPR, and HIPAA compliance.

Our stance

Synthetic media isn't going away. Detection won't “solve” it — but good detection, widely available, keeps most people safe from most deepfake attacks most of the time. We're investing in that reality.