← Back to Daily Builds

Daily build note · October 30, 2026

Real or Scam?

Practice spotting AI-era scams: 12 made-up texts, emails, voicemails and calls (cloned grandkid voice, fake bank, package and toll alerts, plus real ones), with the tells explained, a score, and a printable family safe-word card.

Personal Utilities Interactive demo live Static HTML/CSS/JS

What shipped

Real or Scam? is a practice quiz for parents and grandparents, built for the new wave of AI scams where a voice can be cloned from a few seconds of video. It walks through 12 made-up messages, one at a time, each styled like what you'd actually see: a text bubble, an email with From and Subject, a voicemail transcript, or a phone call. Eight are scams (a cloned "grandkid" calling from jail, a "new number, Mom" text, a bank fraud alert followed by a caller asking for your one-time code, a package redelivery fee, an unpaid toll, a pastor's quiet gift-card favor, a tech-support virus call, and a Social Security arrest threat paid by Bitcoin ATM). Four are real (a pharmacy pickup, a family call that uses the safe word, a bank statement notice, a clinic reminder). Tap Real or Scam and the suspicious phrases get highlighted, with tell chips like Urgency, Secrecy, Strange payment, Spoofed sender and Asks for codes, plus a plain-English explanation. Real messages show why they're OK. At the end you get a score, the ones you missed, and five golden rules. Then the Family Safe-Word Plan builder turns it into a wallet card: a safe word (with a Suggest one button), a backup question, up to three call-back numbers you already know, and house rules. You can print it or copy it. Everything is big text and big buttons, and it works on a phone. The plan saves only on the device, under real-or-scam-v1.

How to try

  • Start with message 1, the "grandkid" voicemail. It sounds like family, but the tells are a rush, a payment app and "don't tell Mom or Dad".
  • Watch for the real ones mixed in. The family call on the saved number that uses the agreed safe word is meant to be trusted.
  • Finish all 12 and look at your score and the misses, then hit Make our Family Safe-Word Plan.
  • Tap Suggest one for a silly safe word, add the numbers you'd call back, then Print card (only the card prints) or Copy as text to send it to the family group chat.

Why this (vs neighbors)

This comes from the Startup Opportunities list ("Norton for AI scams": protect parents from fake voices and emails). Before you sell protection software, the cheapest thing that actually helps is practice plus a plan the family agrees on ahead of time. Earlier builds like Porch Light and Tune-Up were checklists for operators. This one is for the people who get the call.

Scope trim

No call or email screening, no audio clips of cloned voices, no reporting to the FTC or the bank, and no shared family accounts. The 12 messages are fixed and don't change from one try to the next.

Limitations

  • This is practice, not protection. Real scams change fast, and a safe word is one check, not proof. Always hang up and call back on a number you already have before sending money.
  • All messages, names, phone numbers and links are made up. The scam patterns are modeled on common real ones.
  • The plan lives in this browser only. Clearing site data erases it, so print or copy the card.

Suggested next steps

  1. Add a "family mode" that texts a new practice message each week and keeps a streak.
  2. Add short audio clips so people can hear how convincing a cloned voice can be, with consent from the voice owner.
  3. Add a one-tap "I got a weird message" checklist that walks through who to call and how to report it.