VerboSaySpeak freely
Sound like you
Mac voice inputYour style
Your ideas, in your own words. +70%: one user's recording rate vs. 80 chars/min typing; no processing or edits.
Recording output
+70
%
Your ideas, in your own words. +70%: one user's recording rate vs. 80 chars/min typing; no processing or edits.
Your thoughts move faster when familiar names and your preferred style stay with you. VerboSay brings those choices into voice input across your Mac. Language support depends on the speech model you select; startup time varies by device.
The words you use, the microphone you choose, and a way to recover your text when an app cannot accept it.
Add names and technical terms to a personal glossary. Supported speech models can use them during recognition.
Choose a preferred microphone and a backup for when that device is offline.
Choose a faithful, polished, or structured AI style, and tailor writing instructions to each app.
If text cannot be inserted safely, review and copy the result from the current session.
Enable the Labs features and a language model to trigger translation or explanation with spoken phrases.
Choose a supported provider and configure speech, language, and vision models separately. Provider charges may apply.
For short, everyday dictation. Proposed, not available yet.
For longer sessions. Proposed, not available yet.
For additional voice time. Proposed, not available yet.
VerboSay is built for macOS 15.6 or later. One supported speech model recognizes up to 30 languages; language support varies by model. The current Fun-ASR-Realtime option covers Chinese, English, Japanese, and Chinese dialects. Public download is not open yet, and Beta access is by invitation.
Voice recognition needs an internet connection and uses a third-party cloud model. Audio and any context needed for the features you enable are sent to the relevant provider. A full VerboSay privacy notice is being prepared for public release.
You can add names and technical terms to a glossary. Supported speech models may use those terms during recognition. With AI enhancement enabled and a compatible model configured, VerboSay can remove empty fillers and accidental repetition while aiming to preserve your meaning; review important text before sending it.
VerboSay aims to place recognized text in the active text field using macOS Accessibility. Some apps and editors behave differently. If insertion cannot be completed safely, the result remains available to review and copy in the current session.
Yes. You can configure a supported provider and select speech, language, and vision models separately. The provider's usage limits, charges, and privacy terms apply. The Free, Pro, and extra-time plans above are proposals and are not available to purchase yet.