FluidVoice

Project Overview

FluidVoice is an open-source macOS dictation app with on-device AI enhancement. Install with Homebrew in one command.

Key Features

• Fluid Intelligence: on-device AI enhancement (smart formatting, context-aware capitalization, post-processing) with zero data leaving your Mac
• Command Mode: control your Mac by voice
• Write Mode: write or rewrite text in any text field
• Live Preview: real-time transcription overlay
• Multiple speech models: Nemotron Speech 3.5, Parakeet Flash, Cohere Transcribe, Apple Speech, Whisper, and more

Target Audience

• macOS users who want to improve productivity with dictation
• Privacy-conscious users who want local voice processing
• Users looking for an alternative to Wispr Flow and similar commercial products

Quick Start

1. Install via Homebrew: brew install --cask fluidvoice
2. Or download manually from GitHub Releases
3. Launch the app and complete language model setup
4. Start using voice dictation

Core Advantages

• Fully local processing, no cloud API, data never leaves your device
• Blazing fast Parakeet implementation with near-zero latency
• Open-source and free (GPLv3)
• Multiple speech model support

System Requirements

• OS: macOS (iOS and Windows coming soon, Linux soon)
• Architectures: Apple Silicon, Intel
• Additional Dependencies: None

Notes

• Fluid Intelligence is a separately maintained private local AI runtime
• Support development via GitHub Sponsors