WhatsApp Scam Analyzer | Free Forensic Triage | LegalAudit

Most romance, investment, recruitment and recovery scams ride on WhatsApp because the platform is end-to-end encrypted and feels personal. That same opacity makes it the hardest channel to verify with friends or with bank fraud teams. The WhatsApp Scam Analyzer fixes that asymmetry. It cross-references your conversation against a 388-entry international scam corpus, runs language-model-based intent analysis, and surfaces the seven classic manipulation patterns scammers use without you needing to forward screenshots to a stranger.

The analyzer handles the official .zip chat export (Settings → More → Export chat → Without media), individual screenshots, OCR-ed images, voice notes and forwarded PDFs/invoices. Nothing is shared with WhatsApp, Meta or any third party. Everything is processed inside our Swiss-hosted forensic pipeline with a WORM audit chain, so if the conversation later becomes evidence in a complaint or civil suit, the analysis is admissible as expert opinion under Swiss CPP art. 184 and EU Daubert-equivalent criteria.

How it works

Export the chat or paste a fragment

Inside WhatsApp open the contact → More → Export chat → Without media. You get a .zip with a .txt transcript. Or just paste the last 20-50 messages directly into Mythos. Voice notes and forwarded images work too.

Describe the context in one line

Tell Mythos who reached out (random match, recruiter, friend's account, etc.) and what they are asking for (investment, gift cards, crypto wallet seed, romance, fees). This conditions the model on the right scam taxonomy.

Receive a forensic verdict

Mythos returns a structured verdict with confidence score, matched scam patterns from the corpus, specific quote citations from your chat, and a recommended next action (block, report, preserve evidence, file complaint).

Generate the dossier (optional)

If the analyst-grade narrative isn't enough — for example you need to give it to a bank, lawyer, or Polizia Postale — generate the forensic PDF dossier with hash-locked exhibits and Swiss chain-of-custody numbering.

What we detect

Pig butchering grooming sequence (warm-up → small win → bigger ask)

Romance scam scripts (military / oil rig / widower personas)

Fake recovery agent re-victimization patterns

Investment broker impersonation (fake CySEC / SEC / FINRA licensing)

Crypto wallet drain phishing inside chat links

Reverse image search on profile pictures for stolen identities

Language model fingerprinting (was the chat partner an LLM bot?)

Cross-reference with 388-entry international scam corpus

Voice note synthesis detection (is the voice cloned?)

Forwarded malicious APK / IPA / PDF inside the chat

Frequently asked questions

Is the analyzer free, and what does it cost to get a court-usable dossier?

Triage is free with five included dossier credits after a 60-second signup. The court-grade PDF dossier with hash-locked exhibits is included in the Pro subscription: CHF 29/month for unlimited dossiers (soft cap 50/month, contact us beyond). No credit card required for triage.

Can WhatsApp or Meta see the messages I analyze?

No. The chat fragment is processed inside our Swiss-hosted forensic pipeline. Nothing is forwarded to WhatsApp, Meta, OpenAI, Anthropic, or any LLM provider's training set. We use Mythos, our own orchestration over open-weight and contracted-no-training-clause models. Full data flow: legalaudit.ch/en/trust/data-flow.

What does the forensic verdict actually contain?

Verdict (GREEN / AMBER / RED with confidence %), matched scam techniques with corpus IDs (SW-PIG-BUTCHERING, SW-ROMANCE-MILITARY etc.), specific message-by-message citations of the manipulation pattern, a plain-language explanation in your language, and 3 concrete next actions. The forensic dossier adds methodology references (NIST SP 800-86, ISO/IEC 27037), SHA-256 hash of every exhibit, and Swiss-formatted complaint template.