NAME

handy - manual page for handy 0.9.5

SYNOPSIS

handy [OPTIONS]

DESCRIPTION

Handy - Speech to Text

OPTIONS

--start-hidden

Start with the main window hidden

--no-tray

Disable the system tray icon

--toggle-transcription

Toggle transcription on/off (sent to running instance)

--toggle-post-process

Toggle transcription with post-processing on/off (sent to running instance)

--cancel

Cancel the current operation (sent to running instance)

--debug

Enable debug mode with verbose logging

-f, --transcribe-file <WAV>

Transcribe this WAV (16 kHz mono) headlessly and exit. Runs the same batch transcription path as the app ??? no mic, no VAD, no download (the model must already be installed)

--model <MODEL>

Model id to load for --transcribe-file (default: the selected model)

--device-index <N>

Hard-select the compute device for --transcribe-file by its registry index (see --list-devices). Omit to use the persisted accelerator setting. transcribe-cpp (whisper-family) models only

--list-devices

List the transcribe-cpp compute devices (with indices) and exit

--list-models

List the available models (with ids) and exit. Pass an id to --model. Honors --json for machine-readable output

--repeat <N>

Repeat the transcription N times (best_ms reports the fastest run)

--json

Emit --transcribe-file results as JSON

-h, --help

Print help