Upload a suspicious video or image, add a photo of the real person, and paste the link where you found it. Files stay on this computer, except when the AI detector is on: then the file goes to the detector for analysis.
We read hidden labels that AI tools leave inside files, run an AI and deepfake detector on the pixels, and compare faces against the real person's photo.
Paste the post link, or the website the clone is promoting. We look up public records: who registered the domain, when, and who hosts it.
A timestamped record with the file's fingerprint (SHA-256) and every finding. A lawyer can use it to ask the platform for the uploader's IP address and identity by court order.
Demo limits: no hidden label does not prove a file is real (re-uploading strips most labels). The AI detector gives probabilities, not proof; when it is on, the file is sent to Sightengine for analysis. Face match uses an open-source model and can be wrong. Uploader IP addresses are held by the platforms and are only released by legal process.