How it works

Input → Analyze → Explain → Verify → Act. An AI-assisted assessment designed to help you think clearly — not to make decisions for you.

1. 1. Input

   Choose an analyzer — message, email, URL, screenshot or QR (preview). Add the sender and claimed brand if you have them. We show exactly what will be processed.

2. 2. Analyze

   Text is normalized and common secrets are redacted. Rule checks look for scam patterns and look-alike domains; AI supplements them when available. Links are never opened. We try to decode QR codes on your device, sending the image for AI assessment only if no valid web link is decoded. Without AI, text, email and links receive labeled rule-only results; image-only checks are refused without using a slot.

3. 3. Explain

   You get a risk score, confidence, threat categories and signal cards that quote the evidence and explain why it matters.

4. 4. Verify

   Compare the claimed identity with observed indicators, then work through a verification checklist using official channels.

5. 5. Act

   Follow recommended safe actions, save the result, or export a printable security report.

The risk score

A 0–100 score blending deterministic rule checks with the AI model's assessment, plus a separate confidence value that reflects evidence quality and agreement. It's an estimate, not a precise measurement.