§37 — Voice input is prepared before it appears ready. Use the state switcher to review the complete setup contract. The mock follows the current Settings shell and composer control language; it is self-contained, supports dark mode, and fetches no assets.

Voice input

Speak naturally in the composer. Your recording is transcribed on this device and is never sent to a speech service.
⌂Private by design. Audio and transcripts stay on this computer.
This Mac is supported
macOS 14.6 · Apple M1 · 16 GB memory
Compatible
MacmacOS 12+ · Apple Silicon (M1 or newer)
WindowsWindows 10 22H2 or 11 · x64 processor
Memory8 GB recommended
Processor4 CPU cores recommended
W
Whisper Base · English
Fast local transcription for short prompts
Download size 142 MiB · installed only for Voice input
Microphone test
Available after the model is downloaded and verified.
Waiting for model
W
Downloading Whisper Base · English
84 MiB of 142 MiB · about 18 seconds remaining
What happens next
OpenCoworker verifies the model, then asks you to test the microphone. You can leave Settings while this downloads.
Not ready
Your device remains compatible
macOS 14.6 · Apple M1 · 16 GB memory
Compatible
✓Voice input is ready. Transcription runs locally on this computer.
W
Whisper Base · English
Installed and verified · 142 MiB
Ready
Microphone test passed
Your microphone and local transcription engine are working.
“This is a test of voice input in OpenCoworker.”
This Mac
macOS 14.6 · Apple M1 · 16 GB memory
Compatible
Composer behavior in this state
Ask OpenCoworker anything…
Voice input needs setup
Before Ready, the muted mic remains focusable and opens this Settings section. After Ready it becomes the normal record/stop control. Open the four composer states →