OpenDuck

Screen-aware, voice-first, local AI tool that stays lean and out of your away until you ask questions. Built for Apple Silicon (M1 or later).

Demo (In App)

OpenDuck1.mp4

Demo (Screen-aware assistance)

OpenDuck2.mp4

Screenshots

Dual Subtitles	Settings
Screen-Aware Vision	Characters

Features

• Real-Time Voice Interaction: Chat with AI hands-free. Interrupt anytime you want like in a real conversation.
• Screen-Aware Vision: Capture screen regions or the full screen with shortcuts and let the AI see what you see.
• Live Transcription & Subtitle: Real-Time showing trascribtion and subtitle so you don't miss anything important.
• Conversation Management: Search, rename, resume, fork or delete any previous chat session from your local disk.
• Portability: Share your characters (Prompt, Avatar, Voice) easily with .openduck files.
• Engineered for Efficiency: Native macOS application built with Rust and Svelte, optimizing memory for AI models.
• Low Latency: Sub-second latency from the moment you stop talking to when the AI begins speaking.
• Flexible Model Support: Built-in MLX-optimized models or connect to Ollama, LM Studio, and OpenAI-compatible endpoints.

Installation

From Releases

1. Go to Releases
2. Download the latest openduck-beta-xxx.dmg and move it to your Applications folder
3. Start OpenDuck from your Applications folder

Technologies Used

• Tauri - Framework for building tiny, blazing fast binaries for all major desktop platforms
• Svelte - Cybernetically enhanced web apps
• Rust - A language empowering everyone to build reliable and efficient software
• MLX - Array framework for machine learning on Apple silicon

Development

Setup

git clone https://github.com/anslwy/openduck.git
cd openduck
./install.sh

Running the App

./start.sh

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.