Lightning-fast, on-device multilingual TTS in 31 languages. Run entirely locally — no cloud, no API calls, no GPU required.
Quick Start
# Install
pip install supertonic
# Generate speech (first run auto-downloads model ~260MB)
supertonic tts "Hello, this is Supertonic speaking!" --output hello.wav
# Or just say it directly (plays through speakers):
supertonic say "Hello, this is Supertonic speaking!"
The CLI is already built into the supertonic package — no extra scripts needed.
CLI Reference
The supertonic CLI has these commands:
| Command | Description |
|---|---|
supertonic tts (or t) |
Generate speech from text, save to file |
supertonic say |
Generate speech and play it directly |
supertonic list-voices (or lv) |
List available voice styles |
supertonic info (or i) |
Show model information |
supertonic download (or d) |
Download model from HuggingFace |
supertonic version (or v) |
Show version information |
supertonic tts — Generate speech to file
supertonic tts "Hello world" -o hello.wav
supertonic tts "Bonjour le monde" --lang fr --voice F1 -o french.wav
supertonic tts "안녕하세요" --lang ko -o korean.wav
supertonic tts "Your text" --steps 10 --speed 1.0 -o high_quality.wav
Options:
| Flag | Default | Description |
|---|---|---|
-o, --output |
required | Output WAV file path |
--voice |
M1 |
Voice style (M1-M5, F1-F5) |
--lang |
na |
Language code (31 languages) |
--steps |
8 |
Quality: 5 (fast) to 12 (high) |
--speed |
1.05 |
Speed: 0.7 (slow) to 2.0 (fast) |
--model |
supertonic-3 |
Model version |
--max-chunk-length |
auto | Characters per chunk (long texts) |
--silence-duration |
0.3 |
Silence between chunks (seconds) |
-v, --verbose |
— | Enable verbose logging |
supertonic say — Generate and play instantly
supertonic say "Hello, welcome to the world!"
supertonic say "This is a female voice" --voice F1 --steps 10
supertonic say "안녕하세요! 반갑습니다." --lang ko
All same options as tts except --output.
supertonic list-voices
supertonic list-voices
Info & Version
supertonic info
supertonic version
# or
supertonic -v
Full CLI help
supertonic --help # All commands
supertonic tts --help # TTS subcommand details
supertonic say --help # Say subcommand details
Python API
from supertonic import TTS
# Initialize (first run downloads model)
tts = TTS(auto_download=True)
# Get a voice style
style = tts.get_voice_style(voice_name="M1")
# Synthesize
text = "Supertonic is lightning fast, on-device TTS."
wav, duration = tts.synthesize(
text=text,
lang="en",
voice_style=style,
total_steps=8,
speed=1.05,
)
# Save output
tts.save_audio(wav, "output.wav")
Language Support (31)
en ko ja ar bg cs da de el es et fi fr hi hr
hu id it lt lv nl pl pt ro ru sk sl sv tr uk
vi — plus na for language-agnostic mode.
Voice Styles
M1-M5 (male), F1-F5 (female). Custom voices via Voice Builder.
Model Cache
- Linux/macOS:
~/.cache/supertonic/ - Windows:
%USERPROFILE%\.cache\supertonic\
Delete cache to force re-download.
Attribution
This skill was automatically generated from:
supertone-inc/supertonic MIT License Generated by git-to-skill on 2026-05-17
Source materials: