Worklog for task "Add browser-based voice synthesis for letters, syllables, and individual words"

Web Speech API Feature on iPhone

When testing text-to-speech on iPhone, an interesting feature became apparent: both Safari and Chrome effectively use the same iOS system voice engine.

For single capital letters, it behaves very poorly for a learning scenario: instead of the letter itself or its normal pronunciation, it reads out a phrase like "Capital letter B".

For letter flashcards, this breaks the mechanics, so before sending the text to TTS, it had to be normalized to lowercase.

Practical takeaway: the visual representation and the string sent to speechSynthesis do not have to match. The letter can remain capitalized on the screen, while its lowercase variant is passed to TTS.