Sently: Scam & Breach Alert

• 

  Sently analyzes a suspicious message and gives a clear verdict, risk signals, and safe next steps.

Inspiration

Last year my mom almost paid a fake electricity bill. The SMS looked real, the link looked real, and she had nobody to ask at 9pm. Every security app we tried was built for American enterprises, full of jargon and fear. We wanted something a parent can actually use: paste the message, get a straight answer.

What it does

Sently checks anything suspicious. Paste a text, a link, an email, a screenshot or a QR code and it tells you in plain words whether it looks like a scam, before you tap or pay. It also watches your email against known breach databases, shows exactly what leaked and where, and walks you through fixing it. A small companion called Pax explains everything calmly, because scared people make bad decisions. And there is a family safe word: a private word you agree on with your family, so when a call sounds exactly like your son asking for money, you have a way to check. It works offline and never leaves the phone. One plan covers five people.

How we built it

SwiftUI and StoreKit 2 on iOS. The backend is a Cloudflare Worker with D1, a breach database proxy, URL reputation checks and Gemini for deeper message analysis. Subscriptions run on RevenueCat: four products (Personal and Family, monthly and annual), one entitlement, and a default offering the paywall reads live. No price is hardcoded anywhere, the paywall shows whatever the store returns for your country.

Challenges we ran into

Honestly, getting ready for App Review took longer than building some of the features. Truthful privacy labels, cutting anything that was not fully working, explaining which checks run on device and which use cloud AI, and localizing the listing into 18 languages. The harder product problem was tone. Scary warnings make people freeze, so making verdicts calm without making them weak took a lot of rewrites.

Accomplishments that we're proud of

Our parents actually use the scam checker without asking us how. The breach map shows real breach names instead of vague warnings. And the family safe word is something we have not seen any big security app ship.

What we learned

Billing is the part nobody warns you about. RevenueCat saved us weeks here: entitlements, receipt validation and localized pricing across 145 countries just worked, so that time went into the protection features instead.

What's next for Sently

Automated data broker removal, breach alerts on the lock screen, turning on the inbox scanning we already built once the Google CASA audit is done, and Android.

Built With

• cloudflare-workers
• gemini
• ios
• revenuecat
• storekit
• swift
• swiftui
• typescript