izwi tts

Generate speech from text.

Synopsis

izwi tts <TEXT> [OPTIONS]

Converts text to speech using a TTS model. Supports multiple output formats, voice selection, and real-time streaming.

Argument	Description
`<TEXT>`	Text to synthesize (use `-` to read from stdin)

Option	Description	Default
`-m, --model <MODEL>`	TTS model to use	`qwen3-tts-0.6b-base`
`-s, --speaker <VOICE>`	Voice/speaker (name or audio path)	`default`
`-o, --output <PATH>`	Output file path	stdout
`-f, --format <FORMAT>`	Audio format: `wav`, `mp3`, `ogg`, `flac`, `aac`	`wav`
`-r, --speed <SPEED>`	Speech speed (0.5-2.0)	`1.0`
`-t, --temperature <TEMP>`	Sampling temperature	`0.7`
`--stream`	Stream output in real-time	—
`-p, --play`	Play audio after generation	—

izwi tts "Hello, world!" --output hello.wav

izwi tts "Hello, world!" --play

izwi tts "Hello, world!" --format mp3 --output hello.mp3

# Slower izwi tts "Speaking slowly" --speed 0.75 --output slow.wav # Faster izwi tts "Speaking quickly" --speed 1.5 --output fast.wav

echo "Text from pipe" | izwi tts - --output piped.wav cat article.txt | izwi tts - --output article.wav

izwi tts "Hello in cloned voice" \\ --model qwen3-tts-0.6b-customvoice \\ --speaker /path/to/reference.wav \\ --output cloned.wav

izwi tts "Hello in designed voice" \\ --model qwen3-tts-0.6b-voicedesign \\ --speaker "A warm, friendly female voice" \\ --output designed.wav

izwi tts "Long text for streaming" --stream --play

Format	Extension	Notes
`wav`	`.wav`	Uncompressed, highest quality
`mp3`	`.mp3`	Compressed, widely compatible
`ogg`	`.ogg`	Open format, good compression
`flac`	`.flac`	Lossless compression
`aac`	`.aac`	High efficiency compression