VERITAS Vision Engine

Detect deepfakes & synthetic video

Drop an MP4 or WebM and use the trim controls to choose exactly which seconds you want analyzed. VERITAS Vision Engine checks for face-swap deepfakes and fully AI-generated video — returning a forensic confidence score and plain-language verdict.

VERITAS Engine onlineMP4 · WebM · MOVDeepfake + AI videoZero storage

Drag your video file here

or tap to browse · MP4, WebM, MOV · max 100 MB

100% in-memory — nothing is stored

VERITAS Vision Engine

Pick a detection mode, drop your video, then drag the timeline to select exactly which clip to analyze.

AI-generated synthetic video
Face-swap & identity deepfakes
GAN & diffusion model artifacts
Only the selected clip is sent to the API. Your video is never stored.

Inside the VERITAS Vision Engine

How video forensics work

How it works

  1. 1Drop your fileDrop an MP4, WebM or MOV. Choose between Deepfake and AI Generated scan — you can switch the mode even after the video is loaded.
  2. 2Trim your clipA timeline appears under the video preview. Drag the green window to select 5–30 seconds. The credit cost updates live as you adjust.
  3. 3Hit AnalyzeOnly your selected segment is submitted for analysis — nothing outside the window is transmitted.
  4. 4Read the verdictYou get an AI probability score, a confidence level, and a breakdown of the sub-signals the engine used.

What we detect

  • Face-swap & identity deepfakes
  • GAN-generated synthetic faces
  • AI-synthesised video (diffusion / NeRF)
  • Temporal and pixel-level artefacts

Pay only for what you analyze

Credits are charged per selected clip length — 1 credit per 5 seconds, rounded up. The cost updates live as you drag the trim window, so you always know exactly what you are paying before you hit Analyze.

5s = 1 credit10s = 2 credits30s = 6 credits