FluidVoice is a macOS voice-to-text dictation application designed for users seeking a fully local, AI-enhanced transcription solution. It offers real-time transcription, smart typing into any application, and advanced features like meeting file transcription and local AI model integration, providing enhanced privacy and offline capabilities for macOS users.

How It Works

FluidVoice utilizes the Parakeet TDT v3 model for real-time Automatic Speech Recognition (ASR). It supports over 25 languages with auto-detection. For AI enhancement, it integrates with cloud providers like OpenAI and Groq, and crucially, supports local AI models via Ollama for complete offline processing. Transcribed text can be directly input into any application via smart typing, with features like a live preview overlay and menu bar integration for quick access.

Quick Start & Requirements

• Install: Download the latest release from GitHub and move the application to the Applications folder.
• Permissions: Grant microphone access and accessibility permissions for typing functionality.
• Configuration: Set a preferred global hotkey and optionally add AI provider API keys.
• Build from Source: Clone the repository, open the FluidVoice.xcodeproj file in Xcode, and build. Dependencies are managed via Swift Package Manager.
• Requirements: macOS 13.0 (Ventura) or later.

Highlighted Details

• Local AI Processing: Integrates with Ollama for running AI models locally, ensuring data privacy and offline functionality.
• Real-time Transcription: Features live preview mode and uses the Parakeet TDT v3 model for immediate transcription.
• Multi-Provider AI: Supports OpenAI, Groq, and custom AI providers alongside local options.
• Enhanced Dictation: Offers global hotkeys, smart typing directly into any app, and menu bar integration.
• File Transcription: Allows uploading and transcribing audio/video files.

Maintenance & Community

Development updates are shared on X (@ALTIC_DEV). Contribution guidelines are planned for future addition. The project relies on user stars for visibility and motivation, as it is offered as free, open-source software.

Licensing & Compatibility

Licensed under the Apache License 2.0. This permissive license allows for commercial use and integration into closed-source projects without significant restrictions.

Limitations & Caveats

Contribution guidelines are not yet established, indicating a nascent stage for external developer involvement. The project is actively under development, with significant AI improvements anticipated, suggesting potential for future breaking changes or evolving features.