Voice-to-text: Wispr Flow And The Dictation Workflow

Voice-to-text used to be a mediocre accessibility feature. It is now a serious productivity tool. I dictated most of this curriculum's first draft.

The tool I use is Wispr Flow. It runs in the menu bar on Mac (also on Windows, iOS, Android). You hold a function key (or the "Hey Flow" wake word, if you're into that), talk, release the key, and the cleaned-up text gets pasted into whatever application you're in.

Pricing as of April 2026: Free tier (Basic) gets you 2,000 words per week, which is fine for evaluation but not for daily use. Pro is $15/month monthly or $12/month billed annually, with unlimited dictation. Free 14-day trial. No on-device processing: everything goes to the cloud, which matters for privacy-sensitive work (see 6.5).

Why this matters for the vibe coding workflow:

Specifications (section 2.2) are long. Talking through a spec is 3-4x faster than typing it. The first draft of a 30-requirement spec takes 15 minutes to dictate, vs. 45 minutes to type. The cleanup pass on the dictated draft is short because you spoke in full sentences.

Prompts to coding agents are conversational. Talking to an agent the way you'd talk to a coworker is more natural than typing at it. You catch yourself thinking out loud, and the thinking-out-loud often contains the requirement you forgot to write down.

Voice notes for follow-up. When you're reviewing agent output and you spot a problem, dictating "this query doesn't handle the case where status is NULL, also the column name is wrong, it's placed_at not created_at" is faster than typing it.

The trade-off: voice-to-text doesn't work well in environments where you can't talk out loud (open offices, public spaces, the bus). For those contexts, you're typing. But for solo work at a desk, dictation is a real advantage.

Other options: macOS built-in dictation is free, on-device, and lower quality. Superwhisper and Voibe are on-device alternatives if cloud dictation doesn't fit your privacy requirements. Apple's built-in dictation on M1+ Macs is also fully on-device.