Qwen3 ASR is a local speech recognition engine powered by the Qwen3 audio model, running via Appleās MLX framework. All processing happens entirely on your Mac - no cloud, no API key required. Qwen3 ASR delivers strong multilingual accuracy with efficient on-device inference on Apple Silicon.
Features
Fully on-device transcription via MLX
Multilingual speech recognition
Optimized for Apple Silicon
No API key or cloud account needed
Bundled with TypeWhisper - install from the plugin manager
Configuration
No API key required. Qwen3 ASR runs entirely on-device.
Setup
Open TypeWhisper Settings > Plugins
Find the Qwen3 ASR plugin and enable it
The model will be downloaded on first use
Select Qwen3 ASR as your transcription engine in Settings or in a profile