speaker_helper.cli module
Command-line interface for speaker-helper.
Module summary
Exposes the library as a terminal tool with four sub-commands:
synth— synthesise text (or stdin) to a.wavfile, offline or streaming, printing the measured duration and real-time factor.voices— list the preset voices of an engine.clone— register a cloned voice from reference audio (default: the bundled ref-malo) and print its id; missing transcripts come fromvocal-helper.eval— evaluate the engine against a dataset and gate on versioned thresholds (exit code0pass /1fail); seespeaker_helper.eval.speak-from— re-voice audio from a YouTube URL, podcast feed, or the microphone (speech-to-speech); seespeaker_helper.sources.serve— run the REST API server (seespeaker_helper.api).
The --clone family of flags works with any command: speaker-helper --clone
synth "Bonjour" synthesises in the cloned ref-malo voice.
Argument parsing uses the standard-library argparse so the CLI has no
third-party dependency of its own.
Usage example
speaker-helper synth "Bonjour le monde." -o hello.wav --engine kokoro
echo "Texte depuis stdin." | speaker-helper synth -o out.wav
speaker-helper voices --engine kokoro
speaker-helper serve --host 0.0.0.0 --port 8080