Deepfake Detector

Deepfake Detector is an AI tool. A detection service for synthetic video, images and cloned voices. It returns a verdict, confidence score and TrustScore, offers a free tier of 50 monthly scans, and exposes the same model through a REST API with SDKs for common languages.

Deepfake Detector is a synthetic media screening service from deepfakedetector.ai. It examines video clips, still images and audio recordings and returns a verdict of Authentic, Likely Synthetic or Inconclusive, paired with a confidence score and a TrustScore on a 0 to 100 scale. Rather than one general model, the platform runs separate detectors for each media type, so lip-sync drift in a clip, generator fingerprints in a photo and sub-millisecond timing anomalies in speech are each checked with an approach built for that substrate. The product is aimed at two broad audiences. Individuals use it to vet suspicious calls and voice messages, a scenario the site illustrates with documented voice-clone fraud cases. Organizations put it in front of finance, KYC, moderation and security workflows, where a spoofed executive call or a fabricated identity document carries real cost. Newsrooms, fact-checkers, insurers, banks, law enforcement and enterprise security teams are listed as existing users. In practice the workflow is short. An account is created, a file is uploaded by drag and drop, URL paste or right-click through the browser extension, and a verdict comes back with the specific signals behind it. Free accounts are capped at 50 detections a month and 2 minute files, while paid tiers raise both limits to 10 minute files and add PDF and CSV export. Results are probabilistic, so an Inconclusive verdict is treated as a signal to supply a cleaner or higher quality copy rather than as a pass. For engineering teams there is a REST endpoint that accepts a file or media URL and returns JSON with verdict, confidence, trust score and named signals, plus SDKs for Python, Node, Go and Ruby and webhook callbacks for longer files. Uploads are purged from primary storage within 60 seconds, and EU data residency is offered on Business plans and above. The closest alternatives are general purpose AI content classifiers, which tend to cover images and text but treat audio and video as an afterthought; Deepfake Detector's differentiator is depth on synthetic voice and video specifically, balanced against a narrower scope and a metered pricing model.