ThinkAloud Notes

This is a big update: everything below runs 100% on your device, as always. FASTER TRANSCRIPTION • A dedicated on-device speech model (NVIDIA Parakeet) now transcribes recordings by default — dramatically faster and more accurate than before. • Live transcription: watch your words appear as you speak, both for new notes and when adding a recording to an existing note. • Speaker labels: multi-voice recordings show who said what. • No more 1-minute limit when adding a recording to a note — speak as long as you like. • Prefer the AI model? A Settings toggle transcribes with Gemma instead — slower, but supports non-English audio. SMARTER, SAFER NOTES • Silent or empty recordings no longer turn into made-up notes — ThinkAloud tells you nothing was heard instead of letting the AI guess. • Hardened on-device AI: recordings, shared text, and note content are sanitized before reaching the model, so content that tries to give it instructions is captured as text — never obeyed. • Cleaner long transcripts: fixed broken word spacing after the first section of long recordings. POLISH & RELIABILITY • Cancelling a recording you're adding to a note now asks for confirmation before deleting it. • Dozens of reliability, performance, and accessibility fixes across the app, including improved screen-reader support, large-text layouts, and reduced-motion handling.