Reading a transcript and following one are different jobs, and until now the app only knew how to do the second.
Reading
Reading and following are now separate modes. Open a note with the player idle and you get an ordinary page: one size, full contrast, nothing chasing you. Press play and it follows the spoken word, growing the current paragraph and pulling the rest back — the behaviour that was always there, now only when it makes sense.
Timestamps hang in the margin and seek when you click them. Same measurements the PDF export already used, so a printed transcript and the screen finally share an anatomy. On a narrow window the stamp moves above its paragraph instead of squeezing the line.
⌘F finds within the open transcript — match count, next and previous, highlighted in place.
Keyboard: Space plays and pauses, ←/→ skip five seconds, ↑/↓ step by paragraph. The measure is a little narrower, closer to the 66 characters a line is comfortable to read.
Dictation
The first dictation after launch is faster. The model now loads while you are still speaking rather than after you release the globe key, which was the one moment the delay was impossible to hide.
Install
Download the DMG, drag VoiceDumps to Applications, open it. macOS will refuse the first launch because the app is signed but not with a paid Apple certificate — open System Settings → Privacy & Security and click Open Anyway.
Two permissions: Microphone, prompted on first recording, and Accessibility for the globe key, which has to be granted by hand in Privacy & Security → Accessibility. No app can grant that automatically.
Apple Silicon only. Everything runs on your Mac — no account, no API key, no network calls.
sha256 26591222e92eb9ffefa801a7716f147b046707e3b027dfa7f4d1c775524147c3