VoiceInk vs Wispr

I now spend more time talking to my laptop than I do typing into it. And because Apple’s transcription is so rubbish I’ve been a happy customer of Wispr Flow for the last six months.

It’s a great product that Just Works.

But it’s also nuts that I’ve been streaming audio up to some (probably US-based) server and using their compute to transcribe it and then having that text streamed back down when I have a pretty powerfull macbook. Transcription is 💯 something which should be doable to a high-quality with local models.

So three weeks ago I switched to VoiceInk. Things I like:

1. It’s Open Source.
2. It uses local models but is still easy to configure and use.
3. My ramblings stay on my own machine.
4. The chap behind it is a “fairly new developer” and this is his “first time collaborating on Github open-source project” (his words). Which I absolutely love, and kinda blows my mind.

Things I dislike rn:

1. It’s slightly slower to respond to keypresses than Wispr, which means I often loose the firstor last word of a dictation.
2. It’s doesn’t throw away “silent” recordings. Which I do a million times a day by accident.
3. I’m worried about feature bloat and the inevitable lack of focus, UX issues and eventual tech debt which comes with that.

Despite that, if you’repaying for Wispr or SuperWhisper or similar, check out VoiceInk.