Start here
The free tool below runs entirely in this browser — no account, nothing stored or transmitted. Redact anything you would not want kept before you type it.
Open the companion tool →Why the image is the wrong question
Synthetic portraits are generated by the million and pass casual inspection. Tools that claim to detect them produce both false negatives and false positives, which means a verdict about a photograph is not something you can act on either way.
What a generated profile cannot do is answer a live call on demand, meet in a place you choose, and remain consistent under ordinary questions over weeks. Those are the checks that hold, and they cost about five minutes.
The protocol
- Ask for a live video call, and pick a time that is inconvenient for a fake but normal for a real person. Refusal once is nothing; refusal as a pattern is the finding.
- Search the photos yourself if you want to, but treat the result as one weak signal rather than proof — reverse-image hits also appear for stock and stolen photos, not only synthetic ones.
- Notice which channel the other person insists on. Anyone who will only use one app, and will not move to a call or a meeting, is telling you which channel they control.
- Slow anything involving money or personal documents until a live conversation and a real-world meeting have happened.
- If something feels off later, revisit the checklist rather than re-examining the pictures.
What the tool does
Walk through a short checklist and get the two verifications worth asking for, plus what a pattern of refusal means. It scores nothing and renders no verdict about any person; it runs in this browser and stores nothing.
Limit first: This is a verification-prompt aid, not a person-vetting service and not an image-forensics tool. It cannot tell you whether a photo is synthetic, and it cannot tell you whether someone is trustworthy.
Open the companion tool →One human next step
If you already sent money, contact your bank or payment app today and report at ReportFraud.ftc.gov. If you met someone who refused every live verification and asked for money, report the profile to the platform as well.
Research log and safety checks
AI-assist path: not used for a verdict. §1.2 UIC-10 forbids presenting image authenticity as fact; this artifact deliberately converts the question into two live verifications the reader controls. §6.10: not a frontier-pace topic.
§0.5 protection result: category is deepfake-media-verification; crisis-adjacent: no; fear/urgency/scarcity toolkit used: no. The protective function is free and the paid feature is convenience only.