That tool I mentioned? Wispr Flow.

I hit the 2,000-word free limit within a day. Immediately purchased it.

As a developer who’s paid for maybe 4 tools in 20 years, I knew within a day this was solving a real problem extremely well.

I almost shared my referral link (one month free), but that would dilute the message. This isn’t about discounts—it’s about exceptional software.

So what makes it exceptional?

The most amazing part of Wispr is that it puts commas and full stops in the right place. You don’t have to worry about explicitly asking it to add punctuation, and of course it doesn’t make any spelling mistakes.

You don’t have to speak loud. Even if you whisper, it understands what you’re saying. (Hence the name, I suppose.)

I’ve been testing it on iPad and Mac. It went a lot more cleanly and quickly on Mac.

But here’s what really matters: Wispr cleans up your speech so well that most of the time, there is nothing for you to do. This solves the input constraint from yesterday’s post—we can speak 3x faster than we type, but voice tools usually need so much cleanup they’re not worth it.

With Wispr, I don’t have to shout. I can actually really whisper. I don’t have to speak unnaturally. I can just speak in my natural way, and it’s quite enjoyable to talk to yourself in a natural way, which isn’t actually shouting.

Here’s a real-world example: I could actually whisper when other people were talking in the same room while my wife was watching video or TV. And I could still whisper and it was able to figure out my voice and what I had to say. The noise isolation is that good.

This is the kind of “it just works” experience that makes a developer actually open their wallet.

But the real magic happens when you combine Wispr with an LLM. More on that tomorrow.