Wispr Flow vs Superwhisper on Mac: privacy, price, and a third option

Where the audio goes

This is the sharpest difference. Wispr Flow's security FAQ states that Flow processes dictated audio in the cloud and stores customer data in the United States. Superwhisper splits each dictation into two stages, a voice model and an optional language model for post-processing, and each can run locally or in the cloud. With a fully local setup its security page says no audio or text ever leaves your machine.[2][7]

So with Superwhisper the setting you choose decides the answer. With cloud models, audio and text go through API calls to providers such as Anthropic, OpenAI, Deepgram and Groq, or to Superwhisper's own cloud models, and custom vocabulary is sent with the audio. Its agreements with providers include zero data retention, which the page says applies to API usage only.[7]

What you pay

Wispr Flow Pro is $15 a month or $144 a year. Superwhisper Pro is cheaper by the year at $84.99, and it has a $249.99 lifetime plan that Wispr Flow does not list. The free tiers are not alike: Superwhisper's allows unlimited local dictation with any local Whisper model, and Wispr Flow's stops at 2,000 words a week on Mac, with a maximum of 5,000.[5][6]

In Superwhisper's free tier the cloud voice models, AI-powered modes and file transcription are Pro features, and new users get 3,000 words of Pro access before it falls back to Free.[6]

Which one for which person

Choose Wispr Flow for the shortest path to working dictation and the widest platform list, including iPhone and Android, plus team features such as SSO. Choose Superwhisper if you want control over where each stage runs, a free local tier, or a one-time price. We cannot rank them on accuracy, and nobody can from public evidence: there is no shared published benchmark.[5][6][7]

What neither one shows you

Dictation fails quietly. A figure comes out as a different figure, and the sentence stays readable, so nothing makes you look again. Automatic marking of the passages where recognition was unsure is not publicly documented for either app on the pages we checked.[2][6][1]

That is the one thing Witness is built around. It recognises speech only on your Mac, with no cloud option. The text lands at your cursor at once. When a passage deserves a second look, a marker lights up in the menu bar and one click opens the flagged passages, the short audio behind each, and the raw transcript so you can see what cleanup changed. It costs €99 once for two Macs, runs on Apple Silicon with macOS 14.4 or newer, and the 13-day trial is the full version with no account.[1]