Storytime

The problem

Parents read the same picture books every night, in the same voice, over and over. That voice becomes part of the memory. But parents travel, work late, or simply can't always be there. Audiobook apps use professional narrators, and text-to-speech sounds robotic. Neither is the voice your child actually knows.

The solution

Storytime lets a parent digitize a physical picture book and generate audio in their own cloned voice in under ten minutes.

1. Photograph the pages and upload them. Apple Vision OCR extracts the text.
2. Record 5–10 seconds of your voice. No special equipment needed.
3. Hit Generate. Storytime clones your voice on-device and produces audio for every page.
4. Hand it to your kid. They tap play and hear the whole book in your voice, page by page.

Features

• Full-bleed pictures with swipe-to-turn and tap-to-play, designed for kids aged 3–8.
• Voice cloning from a short reference clip.
• Auto-play: audio ends, a 600 ms pause, then the next page plays.
• OCR for English, Simplified and Traditional Chinese, and more.
• No accounts, no subscriptions, no data leaving your home.

How it's built

• Backend: FastAPI. No database; books are JSON files and media lives on disk. Audio generation streams progress page by page.
• OCR: Apple Vision runs on-device with no API key. For Chinese books, a post-processing pass strips the spaces OCR inserts between characters, which otherwise cause unnatural pauses.
• Voice cloning: Qwen3-TTS on Apple Silicon via MLX. All inference happens locally.
• Reader: vanilla JS with large tap targets for small hands.

The hardest part

Voice-cloning quality depends entirely on the reference clip. Too short and the model loses prosody; too noisy and it picks up the room. The sweet spot, 5–10 seconds of clean speech with an exact transcript, took a lot of trial and error.