WhisperFlow

WhisperFlow is dictation that works everywhere. It sits over whatever app you are in: hold a key, talk, and cleaned-up text lands in the field you were focused on. The auto-editing matters most: it strips filler and false starts and punctuates as it goes, so the output reads closer to written prose than to a raw transcript. Real time, and over a hundred languages.

It is in daily rotation here. We use it for first drafts — prompts, pull request descriptions, long Slack replies. Speaking beats typing for anything longer than a paragraph, and the edit afterwards is short because the output arrives tidy. It is worse than typing for anything dense with code, product names or precise punctuation.

Superwhisper is the alternative if you want models running locally rather than through a cloud service, which is the right call when the audio is sensitive. The dictation built into macOS and Windows is fine for short bursts; the gap opens up on long-form. Dragon is the heavyweight for professions needing trained vocabulary and verbatim accuracy. Otter and Granola look adjacent but do a different job — they transcribe meetings you sit in, not text you compose. Pick WhisperFlow for a dictation layer across every app. If you want meeting notes, this is the wrong tool entirely.