-
ScamSpotter — prize scam detected: Likely Scam with matched red flags
Elevator pitch
Paste the suspicious message. Know in seconds if it's a scam. ScamSpotter checks any SMS, email, WhatsApp or DM against the red-flag patterns real scams use — and tells you exactly what to do next.
Inspiration
Everyone has received the "your package could not be delivered" text or the "you won $5,000" message. Most people know scams exist, yet they still work — because they rush you. The gap is not awareness; it is a 10-second pause with clear evidence. ScamSpotter is that pause: paste, see the flags, act with a clear head.
What it does
Paste a suspicious message, or tap one of 4 samples (fake delivery SMS, prize scam, bank phishing, legit message). The rule-based engine scores the text 0–100 across red-flag patterns — fake urgency, prize bait, impersonation, password phishing, suspicious links, link shorteners, insecure http links, payment pressure (gift cards, crypto, wire transfer), too-good-to-be-true money claims, and threats/secrecy — plus lookalike-domain detection (e.g. paypa1-secure.com) and combo bonuses for classic scam signatures (prize + urgency, phishing + urgency, payment + secrecy). Every matched flag shows its name, a plain-English explanation, and the exact phrase that matched. The verdict is Likely Scam, Suspicious, or Looks OK, followed by a "What to do" checklist: don't click, don't reply, verify through the official app or a number you type yourself, report and block. An optional Gemini second opinion is available with the user's own key.
How we built it
Single-file HTML + CSS + vanilla JavaScript. Zero install, no backend, no CDNs. The check runs entirely in the browser, so nothing you paste ever leaves the device. The optional Gemini second opinion uses a key stored only in the browser's localStorage, with silent fallback to the rule-based result on any error.
Challenges we ran into
The hardest part was honesty: a scam detector that cries wolf on every normal message is useless, and one that claims certainty is dangerous. We tuned the patterns so a normal delivery update scores clean while the three scam samples all land on Likely Scam — and made the disclaimer impossible to miss: patterns are flagged, safety is never guaranteed.
Accomplishments that we're proud of
A tool anyone can use in 10 seconds, with zero setup, that turns a moment of panic into a moment of clarity — and shows its work, so users learn the patterns instead of just trusting a score.
What we learned
In fraud prevention, explanation beats automation. Showing why a message was flagged teaches people to spot the next scam themselves, long after they close the tab.
What's next for ScamSpotter
More patterns (job scams, romance scams, crypto giveaways), multilingual pattern packs, and a shareable "check this message" link that never uploads the message itself.
Built with
HTML, CSS, JavaScript, Google Gemini
Built With
- css
- google-gemini
- html
- javascript
