Real or Replay
Consumer app where anyone can paste a video link or upload a clip and get a plain-English read on whether the person on screen looks live, AI-generated, or a deepfake replay.
Consumers who receive suspicious videos from recruiters, online sellers, dating matches, or distant relatives
- Upload or link a short clip and get a 'looks real / suspect / likely fake' call with a one-sentence reason
- Heatmap view that highlights which facial regions tripped the check, like eyes, mouth, or face edges
- Shareable report link you can send to a friend, employer, or platform trust team with the verdict attached
Cheap GenAI faces and voice clones are now routine in job scams, romance scams, and fake seller profiles, and ordinary people have no quick way to check before they wire money.
Deepfake scams surged 250% in 2025 per Cybersecurity Ventures, $12B in global deepfake fraud, $929M in romance scam losses in 2025, plus a real $25M Hong Kong video-call impersonation case — concrete consumer harm.How to Detect Deepfakes in 2025: The Most Reliable Tools Ranked ↗Deepfake Statistics 2025: AI Fraud Data & Trends ↗
Crowded field: Bitdefender RealCheck, Imagera, amige, Scam.ai, Hive, Sensity, Reality Defender, Vastav.AI, Intel FakeCatcher, WeVerify already serve consumers/individuals; the 'paste-a-link + plain English verdict' UX is only mildly differentiated.Bitdefender RealCheck - AI Video Detector ↗amige — AI video detector ↗
Most competitors lean freemium/credits (Imagera 20 credits/scan, Reality Defender 50 free/month, Scam.ai free, Hive no free tier); consumer willingness to pay for trust tools remains unproven, limiting per-user revenue.Imagera AI Video Detector pricing ↗8 Best Deepfake Detectors in 2026, Ranked ↗
EU AI Act and U.S. DEFIANCE Act mandate detection/labeling, Content Credentials from big tech, and deepfake generation keeps improving — long secular tailwind for at least 5–10 years.How to Detect Deepfakes in 2025: The Most Reliable Tools Ranked ↗
Open-source detectors (e.g., FAS-R1 family) lose 45–50% accuracy on in-the-wild deepfakes per Deepfake-Eval-2024; social-platform compression further degrades signal — building a reliable consumer-grade detector is technically hard and false positives can destroy trust.The Best AI Deepfake Detectors in 2026 (Tested on Real Fakes) ↗amige — AI video detector ↗