SFlow
SFlow is a free, open-source macOS dictation app under MIT by Daniel Carreon. It records with a hotkey, sends the audio to Groq's hosted Whisper for transcription, and auto-pastes the text at the cursor through a floating pill. Groq charges about two cents an hour, the README says, so it is paid per use. One caveat: SFlow is online only and sends audio to Groq, so there is no offline mode.
What does SFlow do?
SFlow describes itself as "Open-source voice-to-text for macOS. Wispr Flow alternative at 99% lower cost." The maintainer's own comparison point for the cost claim is a $15 per month Wispr Flow subscription.
The flow is: hold a hotkey, speak, release. The audio goes to Groq's Whisper Large V3 Turbo endpoint and the transcript comes back. A small floating pill shows the state and auto-pastes the text into whatever field is active. The README says it works in any macOS app that accepts text.
Key facts
- Licence: MIT.
- Platform: macOS 15 or later.
- Speech model: Groq's hosted
whisper-large-v3-turbo. - Cost: the README quotes about $0.02 per hour and about $0.60 per month under heavy daily use, at Groq's list price for the Whisper Large V3 Turbo endpoint.
- Internet required: yes. SFlow sends audio to Groq, so there is no offline mode.
- UI: a floating pill that shows the state and pastes the text at the cursor.
- Install: clone the repo, install portaudio with Homebrew, make a Python 3.12 virtual environment, install dependencies, build the app bundle and use
dittoto install it into Applications. - Setup: on first launch, add a Groq API key.
What does it replace, and where does it fall short?
SFlow replaces Wispr Flow for Mac users who want a floating pill and auto-paste and are happy to pay Groq per use instead of a subscription.
Where it falls short: online only. SFlow sends audio to Groq, which is the opposite of Wispr Flow's selling point that the model runs on the device. There is no offline mode and no fallback if Groq is unreachable. The build is from source, not a disk image, so it needs Python 3.12, Homebrew and portaudio set up on the Mac. The project is macOS 15 or later only, and Windows and Linux are not supported. The README names no accuracy figure, latency measurement or language list beyond what Groq's own page lists.
How does SFlow run?
From the README's "Installation" section: clone the repository, run brew install portaudio, create a Python 3.12 virtual environment with uv or python -m venv, install dependencies, run pyinstaller to build the .app, and copy the build into Applications with ditto.
Grant Accessibility, Microphone and Input Monitoring on first launch, then paste a Groq API key into the setup prompt. Hold the default shortcut to speak, release to paste. Each recording is sent to Groq for transcription, so an internet connection is needed every time.
Who is SFlow for?
A Mac developer who already has Homebrew and Python 3.12 ready, who wants Wispr Flow's floating pill and auto-paste without a subscription, and who is happy to pay Groq per use and have audio leave the Mac. Someone who wants everything on the device should look at Handy, Voz, OpenDictate or Dicty instead.
What limits does the README state?
- macOS 15 or later is required.
- Accessibility, Microphone and Input Monitoring permissions are needed.
- An internet connection is required for every transcription; there is no offline mode.
- A Groq API key is required and set on first launch.
- No installer is provided; the app is built from source.
Questions people ask
Is SFlow open source?
Yes. The repository is MIT licensed. The README says SFlow is free and open source and that users can read, fork and extend the code. The Groq service it calls is separate, has its own terms and is paid.
Does SFlow work offline?
No. SFlow sends audio to Groq's Whisper endpoint for every transcription and the README says SFlow needs an internet connection. If a reader wants dictation that works offline, Handy, Voz and OpenDictate run the model on the device.
Can SFlow be used commercially?
Yes for the code, which is MIT. The transcription service is Groq's, and a commercial user pays Groq by the hour of audio. The README gives the Groq list price of about two cents an hour at the time it was written, and notes the price is subject to change on Groq's side.
How does SFlow compare with Wispr Flow?
SFlow is a thinner piece of software that hands transcription off to a hosted service. Wispr Flow's selling point is that the model runs on the device and the user pays a subscription; SFlow pays Groq per use and sends the audio over the network. The open-source alternatives to Wispr Flow page sets SFlow beside Handy, Voz and VoiceInk, which run the model locally.
How much does SFlow cost to use?
The code is free. Groq's Whisper Large V3 Turbo endpoint is the paid part. The README says it costs about $0.02 per hour of audio and about $0.60 per month for heavy daily use, at Groq's list price when the README was written. Groq's current price is on its pricing page; always check it before relying on the figure.
Sources
- SFlow README and LICENSE: github.com/daniel-carreon/sflow, read 2026-11-09.
More voice dictation apps
A free, offline push-to-talk dictation app for macOS, Windows and Linux that runs Whisper or NVIDIA Parakeet on your own machine and pastes the text at your cursor, under MIT.
A GPL-3.0 macOS dictation app that runs Nemotron, Parakeet, Cohere Transcribe, Apple Speech or Whisper on the device, with command and write modes, per-app prompts, and an optional closed local enhancement model.
A dictation and meeting-notes app under MIT that runs Whisper or Parakeet locally or cloud models with your own key, detects Zoom, Teams and FaceTime calls, and labels speakers on device.
A macOS dictation app under GPL-3.0 whose README names Superwhisper and Wispr Flow as the tools it replaces, with local models, per-app modes, a personal dictionary and an AI assistant mode.
A speech-to-text app inside the Epicenter monorepo under AGPL-3.0 that records, transcribes with a provider you choose, cloud or local, optionally polishes the text, and pastes it at the cursor.
An Apache-2.0 macOS dictation app that calls Gemini 3.5 Transcribe with your own key, writes audio to disk from the first millisecond for crash safety, and inserts polished text at your cursor without an account.
Added November 9, 2026. Every claim above comes from the project's README, LICENSE or model card, read on November 9, 2026, or from the GitHub API on the date shown in the panel. Found an error? Write to reveneau@licheo.com and it is fixed in the next weekly pass. Repository: github.com/daniel-carreon/sflow.
Want SFlow running on your own servers?
Reveneau sets it up on your hardware, connects it to your product, and builds the software around it.