Recording
Echosy captures both system audio (any sound playing on your Mac) and microphone audio simultaneously. The recording is transcribed in real time using your selected ASR model.
| Feature | Tier |
|---|---|
| Start / stop recording(⌘+Shift+R) | free |
| Real-time waveform visualization | free |
| Recording timer (elapsed / max) | free |
| Microphone toggle (system audio + mic) | free |
| Save WAV file to workspace | free |
| ASR language selection (24+ languages) | free |
| ASR model switching (Qwen3-ASR, Whisper) | free |
| Extended recording (up to 4 hours)(Free: 15 min max) | pro |
| Auto-summarize when recording stops | pro |
How It Works
- 1Press ⌘+Shift+R (or click Record) to begin capturing.
- 2System audio is captured via macOS ScreenCaptureKit; mic audio goes through the selected input device.
- 3Audio chunks are sent to the on-device ASR model for real-time transcription.
- 4Each segment appears in the Transcript tab with a timestamp and source label (system or mic).
- 5Press ⌘+Shift+R again to stop. The WAV file and transcript are saved to your workspace folder.
Toggle the microphone button to include or exclude your own voice. When disabled, only system audio is captured — perfect for recording meetings where you just want to listen.
ASR Language & Model
Select the transcription language and model from the dropdowns in the recording bar. Echosy supports 10 ASR models across three backends:
| Backend | Models | Notes |
|---|---|---|
| Qwen3-ASR | 0.6B, 1.7B | Best for Chinese dialects + English. Fast on any Mac. |
| MLX Whisper | Small, Large Turbo, Large v3 | Apple Silicon optimized. 99 languages. |
| Whisper | Tiny, Base, Small, Medium, Large | Works on Intel + Apple Silicon. 99 languages. |