FluidVoice
SHA-256FluidVoice is a macOS offline dictation app that converts voice to text entirely on-device, supporting multiple speech models with near-instant transcription. Privacy-first and open-source.
Smart Download
Download Download Version
v1.6.1 · 27.1 MB
Fastest macOS offline dictation app with local AI, supporting multiple models and near-zero latency.
Core Features
- Offline dictation with no cloud dependency
- Multiple speech models: Nemotron, Parakeet, Cohere, Whisper, Apple Speech
- Fluid Intelligence: on-device AI for smart formatting and post-processing
- Command Mode: control your Mac via voice (launch apps, shortcuts, system actions)
- Write Mode: dictate or rewrite text in any app's text field
What It Can't Do
- •Fluid Intelligence is a closed-source private module; core dictation is fully open-source under GPL-3.0
- •Some models (e.g., Cohere, Whisper Large) require large downloads (1.4–2.9 GB)
- •Command Mode relies on accessibility APIs; some system actions may not work
- •macOS only (Apple Silicon + Intel); Windows/iOS versions are planned but not yet available
- •Using cloud AI providers (OpenAI, Groq) sends data externally; opt-in only
Use Cases
- Dictation and note-taking without touching the keyboard
- Voice control of macOS: open apps, run shortcuts, automate workflows
Detailed Introduction
FluidVoice is a fully open-source macOS dictation app that runs entirely offline, featuring the fastest native Parakeet implementation on Mac with near-instant transcription. Unlike cloud-based alternatives like Dragon Naturally Speaking or Google Docs voice typing, FluidVoice keeps all voice and text data on your machine, requiring no API keys or internet connection. It supports multiple state-of-the-art speech models (Nemotron, Parakeet, Cohere, Whisper, Apple Speech) and offers optional local AI enhancement via Fluid Intelligence for smart formatting and context-aware capitalization. The app includes Command Mode for voice control of your Mac, Write Mode for inline text editing, live preview, global hotkey, and per-app configuration. With zero data leaving your device, FluidVoice provides a privacy-first dictation experience that rivals commercial solutions in speed and accuracy.
Tags
Getting Started
Install the software
Open the downloaded dmg file, then drag the app to Applications
Install via Homebrew: brew install --cask fluidvoice, or download from GitHub releases
Grant microphone and accessibility permissions when prompted
Set a global hotkey and choose a speech model; optionally download Fluid Intelligence for local AI enhancement
- Install via Homebrew: brew install --cask fluidvoice, or download from GitHub releases
- Grant microphone and accessibility permissions when prompted
- Set a global hotkey and choose a speech model; optionally download Fluid Intelligence for local AI enhancement
SHA-256 checksum verified
Checksum extracted from GitHub official Release page
SHA256 Checksum
02fcc860592caee537b3d103b72b2eca6933de3b8b49432997298d67366604a6This checksum is extracted from the GitHub Release page. Verify file integrity after download.
All SHA-256 checksums on this platform are extracted from the project's official GitHub Release page, without any modification. You can independently verify them on the GitHub Releases page.
Open Source Transparency
View GitHub SourceUninstall Info
Run 'brew uninstall --cask fluidvoice' in terminal, or drag the app to Trash. Remove residual files from ~/Library/Application Support/FluidVoice and ~/Library/Preferences/com.fluidvoice.plist.
No Extra Dependencies
Ready to use after download. No additional runtime required.
Similar Projects
airi
A self-hosted, open-source AI companion inspired by Neuro-sama, capable of realtime voice chat and playing games like Minecraft and Factorio. Fully cross-platform (Web/Mac/Windows).
Khoj
Your AI second brain – open-source personal AI that works on-device or in the cloud.
ollama
Ollama lets you download, run, and manage large language models locally. One command, multiple platforms, endless possibilities.