The app uses DISTAR phonics, which is a method from the 60s by Siegfried Engelmann. The core idea is simple: teach sounds before letter names, then blend sounds into words, then words into sentences. There's a specific sequence that builds on itself.
On top of that I added spaced repetition, same algorithm as Anki. Cards come back at increasing intervals depending on how well the kid does. This keeps sessions short, maybe 5-10 minutes, and the retention is noticeably better than just drilling the same set of cards every day.
Each card has an AI-generated illustration and spoken audio. All 721 cards are pre-generated and bundled with the app. Nothing calls an API. The whole thing works offline.
What's in it: 721 cards (letters, digraphs, words, sentences), speech recognition for pronunciation practice, no account, no data collection, no ads. Source is on GitHub.
She's been using it for a bit more than 3 months now, roughly 10 minutes a day. She can now recognize phonetic patterns in words and read simple books in English.
No monetization, no data tracking whatsoever. Also available for free on the App Store (if you don't want to deal with the hassle of compiling the app from source): https://apps.apple.com/us/app/early-reader-preschool/id67571...