These names belong to three different things. Whisper is OpenAI's open speech recognition model, published on GitHub. Apps built on Whisper, such as MacWhisper, Superwhisper and WhisperKit-based apps, put that model inside a Mac app. Wispr Flow, spelled without the "h" and the "e", is a commercial dictation service whose own help centre says dictation is transcribed in the cloud. If you typed "whisper flow" or "whisperflow", you were most likely looking for Wispr Flow.
Publisher note: SuprFlow publishes this guide and is one of the products discussed. Competitor details come from each company's own pages, checked on 1 October 2026. Prices and features change; confirm before you buy.
Which is which?
As of 1 October 2026, from each maker's own pages.
| Name you see | What it is | Who makes it | Where speech is turned into text | Correct spelling |
|---|---|---|---|---|
| Whisper | An open speech recognition model with code you run yourself | OpenAI | Wherever you run it, such as your own Mac | Whisper |
| MacWhisper | A Mac app, free and Pro, for transcribing files, meetings and dictation | Good Snooze | On your Mac by default; optional cloud providers | MacWhisper |
| Superwhisper | A dictation app with local and cloud models | SuperUltra, Inc. (privacy policy) | Your choice per mode: local Whisper and other local models, or cloud models | Superwhisper |
| WhisperKit | An open-source Swift framework that apps build in | Argmax | On the device, inside the app that uses it | WhisperKit |
| Wispr Flow | A cloud dictation service | Wispr | In the cloud; an internet connection is required | Wispr Flow |
| "Whisper Flow", "whisperflow" | A common misspelling | — | — | Usually means Wispr Flow |
What is Whisper?
OpenAI's Whisper repository describes it as "a general-purpose speech recognition model" that can do multilingual speech recognition, speech translation and language identification. Its code and model weights are released under the MIT licence.
Whisper is a model, not a dictation app. The README installs it with Python and pip, needs the ffmpeg audio tool, and shows it transcribing audio files from the command line (whisper audio.flac audio.mp3 audio.wav --model turbo) or from Python. It lists model sizes from tiny to large, plus turbo. There is no menu-bar icon or shortcut that types into your other apps. How to run Whisper on a Mac walks through it, and Whisper model sizes explains the sizes.
Because you run it yourself, Whisper processes audio wherever you install it. On your Mac, that means on your Mac.
What are apps built on Whisper?
Many Mac apps package Whisper so you do not have to touch a terminal. Three examples, each described from its own pages:
- MacWhisper. Its home page says it "transcribes audio, video, meetings, and system audio locally on your Mac using Whisper, Parakeet and other AI models", and lists real-time dictation. Its privacy article says transcription is performed locally by default, and that cloud transcription providers, DeepL translation and AI prompts send data out when you turn them on.
- Superwhisper. Its voice model page lists Whisper models through whisper.cpp, from Tiny to Large v3 Turbo, that "run entirely on your device", alongside other local models and cloud models such as S1-Voice, Deepgram Nova and ElevenLabs Scribe.
- WhisperKit-based apps. WhisperKit is Argmax's MIT-licensed framework "for speech-to-text with OpenAI Whisper" on Apple devices. Developers build it into their apps. SuprFlow, which publishes this guide, is one: it recognises speech on the Mac with a downloaded WhisperKit model. What is WhisperKit? explains more.
Two cautions. First, "built on Whisper" tells you which model can run, not which one is running: several of these apps also offer non-Whisper models or cloud providers, so check the setting. Second, an app with "whisper" in its name is not automatically local. Read its privacy page.
What is Wispr Flow?
Wispr Flow is a dictation service from Wispr for Mac, Windows, iPhone and Android. Its help centre says Flow transcribes your words, cleans up the result and inserts it into the text field you are using, and that "An internet connection is required for transcription." Another help article says "Dictation is transcribed in the cloud", and Wispr's Enterprise Privacy page says it does not offer on-premise, private-tenant or offline deployment.
The Wispr pages we read do not name the speech model behind Flow, so we do not say whether it uses Whisper or anything else. The similar name does not tell you either way. What matters for privacy is what Wispr does document: the audio goes to its cloud. Is Wispr Flow safe? covers its training, storage and audit details, and Does Wispr Flow work offline? covers the connection question.
Which ones run locally on your Mac?
| Runs on your Mac? | According to | |
|---|---|---|
| Whisper | Yes, when you install and run it on your Mac | OpenAI's README |
| MacWhisper | Yes by default; cloud providers are optional | MacWhisper privacy article |
| Superwhisper | Yes with local models; cloud models are a per-mode choice | Superwhisper voice models |
| WhisperKit apps | WhisperKit itself runs on the device; check each app for other network features | argmax-oss-swift README |
| Wispr Flow | No; transcription happens in the cloud | Wispr help centre |
"Runs locally" means speech is recognised on the Mac. It does not mean the app never uses the network: licence checks, model downloads, updates and optional AI features can all go online. On-device dictation vs no internet explains the difference.
Why does the spelling matter?
Because you may download the wrong thing. Search for "whisper flow" and you can land on a Whisper-based app when you wanted Wispr's cloud service, or the other way round, and the two handle your audio differently. Lookalike download sites are a real risk too: MacWhisper's own site warns that "there are scam websites online that pretend to be the official website for MacWhisper, and they may contain malware downloads."
Before installing, check the address: wisprflow.ai for Wispr Flow, macwhisper.com for MacWhisper, superwhisper.com for Superwhisper, and github.com/openai/whisper for OpenAI's model. How to check a signed Mac release shows how to see who signed the app you installed.
Which should you use?
- You want dictation that types into any app and you are always online: Wispr Flow is built for that. See Wispr Flow pricing.
- You want to transcribe recordings or meetings on the Mac: MacWhisper is built around files and meetings.
- You want to experiment with the model itself: install OpenAI's Whisper.
- You want push-to-talk dictation with speech recognised on the Mac: a Whisper-based dictation app. SuprFlow is one paid option: hold Fn, speak, release, and the text is pasted at the cursor, with local cleanup of punctuation, capitalization and numbers. It costs $125 once, is Mac-only (Apple Silicon, macOS 14 or later), is not open source, and has no free tier, file transcription or meeting recorder. Free local options exist too; Best offline dictation app for Mac compares them.
We have not run a controlled accuracy benchmark of any of these.
Questions people ask
Is Wispr Flow the same as Whisper?
No. Whisper is OpenAI's open speech model. Wispr Flow is a separate company's cloud dictation service. Wispr's pages we read do not say which speech model it uses.
Is it Whisper Flow or Wispr Flow?
The product is spelled Wispr Flow. "Whisper Flow" and "whisperflow" are common misspellings of it.
Is MacWhisper the same as Wispr Flow?
No. MacWhisper is a Mac app from Good Snooze that transcribes locally by default using models such as Whisper and Parakeet. Wispr Flow transcribes in Wispr's cloud.
Does Wispr Flow work offline?
No. Wispr's help centre says an internet connection is required for transcription.
Is OpenAI Whisper free?
Its code and model weights are released under the MIT licence, so you can download and run it without paying OpenAI. Apps that package it set their own prices.
Related guides
- Is Wispr Flow safe?
- MacWhisper vs Wispr Flow
- Superwhisper vs Wispr Flow
- What is WhisperKit?
- How to run Whisper on a Mac
- SuprFlow vs Wispr Flow
Sources
- OpenAI — whisper GitHub repository and README
- MacWhisper — official site
- MacWhisper Support — Keeping Transcriptions Private
- Superwhisper Docs — Voice models
- Superwhisper — Privacy Policy
- Argmax — argmax-oss-swift (WhisperKit) GitHub repository
- Wispr Flow Help Center — What is Wispr Flow?
- Wispr Flow Help Center — What to expect from Flow: accuracy and known limitations
- Wispr Flow — Enterprise Privacy