Voice got a massive upgrade

Better, faster, stronger (but no vocoder 🤖)

We forgot to hit the send button last week 😅

So it you want to catchup on last week new features, it is here:

Save as DOCX, Instant URL Opening, and Rich Markdown , our new Quick ActionsOlivier

Hey everyone,

We’ve overhauled our entire voice system to make dictation smoother, faster, and more reliable, especially when it comes to capturing your very first word.

Oh and Facetime works perfeclty now :)

What We Did

We rebuilt the whole voice pipeline from scratch to improve responsiveness and quality.

The Results

Instant capture from the first millisecond — Your very first word is recorded the moment you trigger the keyboard shortcut, no lag or missed beginnings.
Seamless input switching — Change your microphone mid-recording without any disruption.
Enhanced Bluetooth stability — Clearer audio and fewer dropouts with wireless devices.
Sharper speaker identification for local models — No more imaginary voices or phantom participants in meetings.

Getting the most of it - our guides

Talk to Your Mac: Alter's Speech-to-Prompt Guide

Real-time Meeting Transcripts with Whisper Pro

🦸The super hero trick

Did you know you can use speech-to-prompt with any selected text?

Just select a text, press and hold the mic icon or your voice shortcut, then ask Alter a question.

It works really well with requests like “explain what this means.”

Your Feedback Matters

Voice is tough tech. If you experience any issues or want specific improvements, please let us know, your voice matters.

Try out the new experience today by updating Alter and start speaking with confidence.

To crisp conversations,

Olivier

Full Changelog

New Features & Enhancements

Dictation: Performance improvements on macOS 14.2+
Transcription: Support for full audio file transcription with local models
Local Diarization: Enhanced speaker detection

Bug Fixes & Stability

Audio: Support for FaceTime calls added
Audio: Improved support for Bluetooth headphones and handling changes during recording
Dictation: Fixed various crashes introduced in beta40
Custom Endpoint: Stopped sending the "thinking" parameter