AudioInk

Local audio transcription application powered by OpenAI's Whisper model. Built with Tauri 2 and Rust for native performance and privacy.

Features

• Local Processing: All transcription happens on your device - your data never leaves your computer
• Multiple Whisper Models: Choose from Tiny, Base, Small, Medium, or Large-v3-Turbo based on your needs
• Audio/Video Support: Transcribe MP3, WAV, M4A, FLAC, OGG, MP4, MOV, MKV files
• YouTube Integration: Fetch captions or transcribe with Whisper using yt-dlp
• Audio Speedup: Optional 1.25x-2.0x acceleration for faster transcription (requires ffmpeg)
• Progressive Display: See transcription results in real-time as chunks complete
• Multi-language: Auto-detect or specify the audio language (English, Spanish, French, German, Portuguese, Japanese, Chinese, and more)
• Timestamps: Optional timestamp markers in transcriptions [HH:MM:SS]
• History: Automatically saves transcriptions with metadata
• Lightweight: ~15MB app size with minimal resource usage

Requirements

macOS

• macOS 10.15+ (Apple Silicon or Intel)
• ~150MB - 1.5GB disk space (depending on chosen Whisper model)

Linux

• Ubuntu 20.04+, Fedora 36+, or equivalent
• ~150MB - 1.5GB disk space (depending on chosen Whisper model)
• Build dependencies: build-essential, cmake, libwebkit2gtk-4.1-dev, libssl-dev, libayatana-appindicator3-dev, librsvg2-dev

Windows

• Windows 10/11
• ~150MB - 1.5GB disk space (depending on chosen Whisper model)

Optional Dependencies

Tool	macOS	Linux	Windows
yt-dlp (YouTube transcription)	brew install yt-dlp	sudo apt install yt-dlp	winget install yt-dlp
ffmpeg (audio speedup)	brew install ffmpeg	sudo apt install ffmpeg	winget install ffmpeg

Installation

From Source

1. Install prerequisites:

   macOS:

   # Install Rust
   curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
   
   # Install Node.js (v18+)
   brew install node

   Linux (Ubuntu/Debian):

   # Install Rust
   curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
   
   # Install system dependencies
   sudo apt update
   sudo apt install -y build-essential cmake libwebkit2gtk-4.1-dev libssl-dev libayatana-appindicator3-dev librsvg2-dev
   
   # Install Node.js (v18+)
   curl -fsSL https://deb.nodesource.com/setup_20.x | sudo -E bash -
   sudo apt install -y nodejs

   Windows:

   # Install Rust from https://rustup.rs
   # Install Node.js from https://nodejs.org
   # Install Visual Studio Build Tools with C++ workload

2. Clone and build:

   git clone https://github.com/fabianhtml/audioink-rs.git
   cd audioink-rs
   npm install
   npm run tauri build

3. The app will be available in src-tauri/target/release/bundle/

   • macOS: .dmg and .app
   • Linux: .deb, .rpm, and .AppImage
   • Windows: .msi and .exe

Development

npm install
npm run tauri dev

Usage

1. Download a Model: Open Settings and download your preferred Whisper model

   • Tiny (75 MB): Fastest, lower accuracy
   • Base (142 MB): Good balance for most uses
   • Small (466 MB): Better accuracy
   • Medium (1.5 GB): High accuracy
   • Turbo (809 MB): Best quality, optimized for speed

2. Transcribe Audio:

   • File Tab: Click to select an audio/video file
   • YouTube Tab: Paste a YouTube URL to fetch captions

3. Options:

   • Select language or use auto-detect
   • Enable timestamps in Settings for time-marked output
   • Enable audio speedup (1.25x-2.0x) for faster processing (note: not compatible with timestamps)

4. Export: Copy to clipboard or save as text file

Architecture

audioink-rs/
├── src/                    # Frontend (Vanilla JS)
│   ├── index.html
│   ├── main.js
│   └── styles.css
└── src-tauri/              # Backend (Rust)
    └── src/
        ├── commands/       # Tauri commands (API)
        ├── core/           # Whisper engine, audio processing
        ├── models/         # Data structures
        ├── persistence/    # History management
        └── utils/          # Error handling, helpers

Tech Stack

• Frontend: Vanilla HTML/CSS/JavaScript
• Backend: Rust + Tauri 2
• Transcription: whisper-rs (Whisper.cpp bindings)
• Audio Decoding: Symphonia

License

MIT

Acknowledgments

• OpenAI Whisper - Speech recognition model
• whisper.cpp - C++ port of Whisper
• Tauri - Desktop app framework