Repolode
Submit
altic-dev

altic-dev/FluidVoice

Open source macOS dictation app that turns speech into text with on-device models and optional AI cleanup, a local alternative to Wispr Flow.

Open-source alternative to:Wispr Flow

View on GitHubWebsite
Stars
12,328
Forks
895
Open issues
131
Last commit
18 hours ago

What is it?

FluidVoice is a voice-to-text dictation app for macOS. You hold a global hotkey, speak, and the transcribed text is typed into whatever app has focus. It supports several speech engines, including NVIDIA Parakeet and Nemotron models, Cohere Transcribe, Apple's built-in speech recognition, and Whisper models that also run on Intel Macs.

On top of transcription, FluidVoice can enhance text with smart formatting and capitalization, using either a local model called Fluid Intelligence or cloud providers such as OpenAI and Groq with your own API key. Command Mode lets you control the Mac by voice, and Write Mode rewrites selected text in any text field. Local audio history with ZIP export is optional.

The app requires macOS 15 or later and microphone and accessibility permissions. Fluid Intelligence is a separate, privately maintained component, while the dictation app itself is GPLv3 open source. Windows, iOS and other platforms are described as planned.

Key features

  • 01Global hotkey starts voice capture from any app and types the result directly
  • 02Multiple speech engines, including Parakeet, Nemotron, Cohere, Apple Speech and Whisper
  • 03Optional AI enhancement via local model or OpenAI, Groq and custom providers
  • 04Command Mode for controlling the Mac by voice
  • 05Write Mode for rewriting selected text in any text field
  • 06Local-first design where audio stays on the Mac unless a cloud provider is enabled

Use cases

  • Dictating emails, documents or chat messages on a Mac without typing
  • Replacing a paid cloud dictation tool with an on-device setup
  • Rewriting or polishing text in any app by voice
  • Dictating in several languages with a choice of speech models

Quick start

  1. 1

    Install with Homebrew

    brew install --cask fluidvoice
  2. 2

    Or download the latest release from the GitHub releases page

  3. 3

    Grant microphone and accessibility permissions when prompted

  4. 4

    Set a global hotkey in settings and complete onboarding to choose a voice model

Things to know

  • macOS only; iOS and Windows are not yet released.
  • Requires macOS 15.0 or later, and most speech models need an Apple Silicon Mac.
  • Fluid Intelligence is a separate privately maintained component, so its on-device enhancement is not open source.

Topics

#speech-to-text#dictation#macos#whisper#on-device-ai#swift#voice-control#ai#ios#llama-cpp

FAQ

What is FluidVoice?
Open source macOS dictation app that turns speech into text with on-device models and optional AI cleanup, a local alternative to Wispr Flow. FluidVoice is a voice-to-text dictation app for macOS. You hold a global hotkey, speak, and the transcribed text is typed into whatever app has focus. It supports several speech engines, including NVIDIA Parakeet and Nemotron models, Cohere Transcribe, Apple's built-in speech recognition, and Whisper models that also run on Intel Macs.
Is FluidVoice free to use? What license does it have?
FluidVoice is open source under the GPL-3.0 license. Check the license terms before using it in commercial products.
Is FluidVoice still maintained?
Actively maintained. The last commit to FluidVoice was on Oct 20, 2026.
What can I use FluidVoice for?
Dictating emails, documents or chat messages on a Mac without typing Replacing a paid cloud dictation tool with an on-device setup Rewriting or polishing text in any app by voice Dictating in several languages with a choice of speech models
How do I install FluidVoice?
Install with Homebrew: brew install --cask fluidvoice Or download the latest release from the GitHub releases page Grant microphone and accessibility permissions when prompted Set a global hotkey in settings and complete onboarding to choose a voice model
Where can I get FluidVoice?
The source code and installation instructions are on GitHub: https://github.com/altic-dev/FluidVoice

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects