ScamCheck Claude - 2026 Pricing, Features, Reviews & Alternatives

Free with your Anthropic API key. Find out whether the message in front of you is a scam, before you reply, click or send money.

ScamCheck Claude v5.2 analyzes photos, screenshots and PDF documents of emails, letters, texts, advertisements and personal communication, and names the fraud for what it is. Powered by Claude Sonnet 5, it goes beyond keyword flagging to read the structure of the deception itself: manufactured urgency, borrowed authority, grooming scripts, and the point money is asked for. It recognizes seventeen fraud families, and is kept current as new ones appear:

Employment and task fraud: pay-to-get-paid work against a promise of reimbursement and commission

Tech support and remote access takeover: a fake security alert that ends with remote software installed and the account drained

Recovery and revictimization: a fee to recover money already lost, sold to known victims

Bank and payment impersonation: the fraud-department callback and the safe-account transfer

Romance and relationship investment fraud: love bombing, grooming toward one platform, then the withdrawal trap

Investment and crypto fraud: fake exchanges, trading bots, wallet drainers, deepfaked endorsements

Extortion and coercion: sextortion including AI imagery, claimed device compromise, virtual kidnapping

Personal impersonation: a cloned voice or deepfaked video of a relative, an executive, or an official

Government and authority impersonation: agencies and courts named or styled by a sender that is neither

Deceptive mailers and legal-lookalike marketing: solar and home improvement lien terms, timeshare exit, notario services

Smishing and account alerts: unpaid tolls, undelivered packages, suspension notices that harvest a card

Invoice, subscription, and payment redirect: fake renewals, business email banking changes, payroll diversion

Prize, inheritance, and windfall: a fee or tax to release money you never earned

Unsolicited insurance and warranty: vehicle service contracts, home warranties, Medicare outreach, ghost broking

Unsolicited credit, loan, and asset purchase: advance fee loans, fake pre-approvals, we-buy-your-car and we-buy-your-house

Unsolicited financial products: timeshares, reverse mortgages, annuity rollovers, pension advances, metals IRAs

Free meal and seminar financial pitches: a free dinner tied to a retirement workshop that sells commissioned products

The last five arrive overwhelmingly as physical mail, which is why a photograph of what came in the envelope is first-class evidence here rather than a secondary case.

Provide your Anthropic API key, then upload a file or capture one with the live camera, and you get a report naming the fraud family, whether the contact was solicited, where in the scheme you stand, and whether money already sent can be recovered, with a Risk Score, a Verdict, and every red flag in plain language.

Solicited versus unsolicited is the strongest single signal on any commercial offer and is decided on every scan: personalization is not a relationship, and a mailer quoting your lender or loan balance bought that from a data broker.

The evidence you submitted appears above the verdict, with red flags, risk score, tokens, and cost as large labelled figures at the top. Reports save as text or PDF, and the app runs in English, German, or any of 100+ languages your browser displays.

A scan costs about four to eight cents for a photo or screenshot, and up to twelve cents for a multi-page PDF, billed by Anthropic directly to your own account with no markup and no subscription. Five dollars in credit covers roughly sixty to over a hundred scans.

Your API credentials and the content you analyze are never held or accessed; data travels only between your browser and Anthropic, encrypted in transit. There is no account to create and nothing to install.

Use is subject to the End User License Agreement, Terms of Service, and Privacy Policy. Requires your own Anthropic API key.