Writing
Tech notes and field-tested insights from building products.
Dev note: turning any text into something you can hear, read, and shadow — the story behind YomiPlay's Text-to-Speech
Got the text but no audio? YomiPlay's new Text-to-Speech turns any passage into natural narration plus synced subtitles — generated on your phone in seconds, ready for the same close-listening, shadowing, and review as any imported audio.
A Markdown editor that adds furigana to Japanese, made for my daughter — YomiNote
A Chinese-speaking fifth-grader living in Japan can read the kanji but can't always pronounce them. YomiNote was built for exactly that.
Meaning without sound — the "translucent" state of Chinese-background kids learning Japanese
Chinese-background kids learning Japanese often fall into a "translucent" state — meanings come through at a glance, but the readings stay hazy. A narrow gap worth building a tool for.
I built an app called YomiPlay — it turns Japanese audio and video into watchable, shadowable, editable, shareable learning material
Auto-transcribe meeting recordings, podcasts, and YouTube videos into clickable, furigana-aware, editable, exportable subtitles. YomiPlay isn't a player — it's a pipeline that turns a one-time listen into a reusable learning asset.
I built a Chrome extension to make Japanese katakana loanwords easier to read — meet YomiMark
For Chinese-speaking developers reading Japanese tech pages, the real pain isn't kanji — it's katakana loanwords. You can read every character but can't tell the English word in disguise. YomiMark quietly maps them back, fully offline, triggered by text selection.