Hold
Press Control.
The global trigger starts capture wherever you are working.
Private voice writing for Mac
Speak. It lands in under 1 second.
Your voice stays on your Mac.
<1s
Sentences, notes, and paragraphs land as text in under a second.
< 1 sec result
Move the product review to Friday. Add the open questions. Send the recap to Maya.
No account.Core dictation works without sign-in.
No card.Voice writing is not gated by billing.
No upload.Your speech stays on device for local transcription.
No waiting.Sentences, notes, and paragraphs return as you speak.
500-word note
Nine minutes back.
40 WPM typing takes 12.5 minutes. 150 WPM speech takes 3.3 minutes. About 9 minutes back on a 500-word note.
250 WPM is elite typing; the comparison above uses normal averages.
Less input time
73%
Typing
12.5 min
Speaking
3.3 min
~9 min
Speak into any app. Get text back in that app. If paste is blocked, the transcript is still copied.
The whole product is the loop.
Scrivora feels like a Mac utility, not a cloud workspace. No account to manage. No dashboard to babysit. No paste path to chase.
Speak
Say the text.
Names, quick replies, and long notes all use the same local path.
Local
Run realtime ASR.
Local models handle speech without a required cloud speech API.
Write
Get text back fast.
Text lands in the focused app, or stays copied when focus changes.
Private means boring by default.
The safest voice app is the one that does not make speech data interesting to a server. Scrivora keeps core dictation local and makes network use explicit.
Captured for transcription on your Mac.
Core dictation does not need to send your voice to a cloud speech API.
Returned to the app you were using.
When insertion is unavailable, the transcript remains on the clipboard instead of becoming an error screen.
Local, recent, and useful.
Recent transcripts stay local and copyable without turning the product into a noisy log viewer.
Downloaded by choice.
Model downloads and app updates use the network only when you choose them. Core dictation does not require a hosted speech account.
Open source. Simple install.
Download the Mac app as a DMG, drag Scrivora into Applications, and start dictating. Clone the source only if you want to inspect, edit, or build it yourself.
DMG for normal users.
Open the downloaded DMG, drag Scrivora.app into Applications, and run it from the menu bar.
Edit it if you want.
The source lives on GitHub for people who want to audit, change, or build Scrivora themselves.
Model credits.
FluidAudio Parakeet V3 is named clearly, with model terms tracked beside the release docs.
Plain answers beat product fog.
Do I need to sign in?
No. Core dictation is account-free.
Do I need a card?
No. The voice-writing loop is not gated by card capture.
Does speech leave my Mac?
Core dictation uses local transcription. Model downloads or update checks are separate network actions.
Where do releases live?
Source tags and release assets live on GitHub. Release notes and update metadata live on scrivora.me. The DMG appears only after the signing checks pass.