# yovoice models and sound fonts

> Models and score sound fonts supported by yovoice v0.5.0. File sizes follow the current model guide and model catalog; GB uses 1e9 bytes. Model and sound font licenses are separate from the yovoice Apache-2.0 software license.

## Speech and music models

| Model | Parameters | Model file size by precision | Core capabilities |
| --- | --- | --- | --- |
| IndexTTS 2.0 | — | Q8 · 3.63 GB<br>F16 · 4.65 GB<br>ORIG · 8.08 GB | Chinese/English voice cloning, emotion control, reference performance |
| IndexTTS 2.5 | — | Q8 · 3.50 GB<br>F16 · 4.55 GB<br>ORIG · 7.89 GB | Multilingual voice cloning, emotion control, pronunciation editing |
| VoxCPM2 | 2B | Q8 · 2.96 GB<br>BF16 · 4.77 GB<br>ORIG · 4.96 GB | Text-guided voice design, voice cloning, transcript-assisted cloning |
| OmniVoice | 0.6B | Q8 · 1.35 GB<br>BF16 · 1.64 GB<br>F16 · 1.64 GB | Attribute-based voice design, voice cloning, non-verbal sound tags |
| Qwen3-TTS Base | 0.6B | Q8 · 1.99 GB<br>BF16 · 2.52 GB | Reference voice cloning, optional transcript guidance, multilingual speech |
| Qwen3-TTS Base | 1.7B | Q8 · 2.70 GB<br>BF16 · 4.20 GB<br>ORIG · 4.54 GB | Reference voice cloning, optional transcript guidance, multilingual speech |
| Qwen3-TTS CustomVoice | 1.7B | Q8 · 2.82 GB<br>BF16 · 4.18 GB | 9 built-in voices, text-guided style and emotion |
| Qwen3-TTS VoiceDesign | 1.7B | Q8 · 2.82 GB<br>BF16 · 4.18 GB | Voice design from natural-language descriptions, no reference audio required |
| Breeze TTS 2 | ~3B | Q4 · 3.79 GB<br>Q8 · 5.08 GB<br>BF16 · 7.34 GB | Chinese and English voice design, cloning and voice direction; GPU recommended, requires the v0.9.1 engine |
| Ming 3.0 TTS | 0.5B | Q8 · 1.74 GB<br>BF16 · 3.00 GB | Voice design and cloning in Chinese, English, Japanese, Korean, German, French and Spanish; 44.1 kHz; requires engine v0.9.3 |
| Voxtral 4B TTS | 4B | BF16 · 8.24 GB | 20 built-in voices in 9 languages; CC BY-NC 4.0; requires engine v0.9.3 |
| Plapre 3 | — | Q8 · 1.45 GB | Danish built-in voices and cloning; requires engine v0.9.3 |
| XTTS v2 | — | Q8 · 1.86 GB | Voice cloning in 17 languages; Coqui Public Model License; requires engine v0.9.3 |
| Gradium Phonon | — | Q8 · 850 MB | Lightweight English speech; default voice only; requires engine v0.9.3 |
| Drama 3 / Drama3-small | — | Q8 · 2.10 GB | Speaker and event tags, default voice and cloning; Fish Audio Community License; requires engine v0.9.3 |
| Kokoro-82M 1.0 Official | 82M | Q8 · 189.55 MB<br>BF16 · 211.95 MB | 49 built-in voices, multilingual speech excluding Japanese; supports voice blending |
| Kokoro-82M 1.1-zh | 82M | Q8 · 255.32 MB | 100 Chinese and 3 English voices; experimental |
| ACE-Step 1.5 Turbo | — | Q8 · 6.19 GB<br>BF16 · 10.09 GB | Songs and instrumentals from a style prompt and lyrics, 50+ vocal languages; requires engine v0.9.1 |
| ACE-Step 1.5 XL Turbo | — | Q8 · 10.71 GB<br>BF16 · 15.28 GB | Larger ACE-Step model for higher-quality songs and instrumentals; requires engine v0.9.1 |
| Stable Audio 3 Small Music | — | Q8 · 1.68 GB<br>F16 · 2.36 GB | Instrumental music from English descriptions, up to 2 minutes; rework and inpainting; Stability AI Community License |
| Stable Audio 3 Small SFX | — | Q8 · 1.68 GB<br>F16 · 2.36 GB | Sound effects (1–30 s) from English descriptions; rework and inpainting; Stability AI Community License |
| Stable Audio 3 Medium | — | Q8 · 3.58 GB<br>F16 · 5.83 GB | Higher-quality music and sound effects up to 6 minutes; rework and inpainting; Stability AI Community License |
| YuE2 3B | 3B | Q4 · 2.93 GB<br>Q8 · 4.53 GB<br>BF16 · 7.53 GB | Full vocal songs from lyrics and style, length follows lyrics, score-based arrangement; requires engine v0.9.1 |

All listed models are available in the App and CLI. Use `yovoice models list --json` for model IDs, available precisions, file sizes, voices, and advanced parameter definitions.

## Score sound fonts

Score projects and `yovoice score render` use General MIDI sound fonts. Rendering needs no inference engine or GPU, and the same score renders the same way every time.

- **MuseScore General** (`musescore-general-sf2`): default sound font with all 128 General MIDI instruments and drum kits; SF2, about 216 MB.
- **GeneralUser GS** (`generaluser-gs-sf2`): optional alternative; about 32 MB.
- **SGM-V2.01** (`sgm-v2-sf2`): optional alternative; about 247 MB.

The CLI can export a mixed WAV, per-track stems, MIDI, and score JSON. See the Score guide for input format, mixing, and license notes.

## Licenses and commercial use

- **Stable Audio 3**: weights use the Stability AI Community License (free for annual revenue below USD 1M).
- **OmniVoice**: weights use CC-BY-NC and are restricted to non-commercial use.
- **YuE2**: weights use CC BY-NC 4.0 and are restricted to non-commercial use.
- **Voxtral 4B TTS**: weights use CC BY-NC 4.0 and are restricted to non-commercial use.
- **Breeze TTS 2**: weights use the BreezeBlue Research and Non-Commercial License. Generated speech is not licensed for commercial use either.
- **Drama 3**: covered by the Fish Audio Community License.
- **XTTS v2**: covered by the Coqui Public Model License.
- **IndexTTS**: covered by a separate model license.
- **Sound fonts**: the Score guide describes GeneralUser GS as free for commercial music and states that the SGM-V2.01 author provides no license. Check each sound font's terms before redistribution or commercial use.

For every other model, check the publisher's current terms. Before commercial use, also make sure you have the necessary rights to reference audio, voices, and other source material.

## Hardware quick guide

| Use case | Recommended hardware | Recommended models |
| --- | --- | --- |
| Lightweight narration | Any CPU / 4 GB+ RAM | Kokoro-82M (Q8 about 189 MB) |
| High-fidelity cloning and voice design | 4–6 GB VRAM / 8 GB+ Apple Silicon | VoxCPM2 (2B, 48 kHz), IndexTTS 2.0/2.5, Qwen3-TTS, Ming 3.0 TTS |
| Full-song, instrumental, and SFX generation | 12 GB+ VRAM / 16 GB+ Apple Silicon | ACE-Step 1.5 Turbo, Stable Audio 3, YuE2 3B |

## Canonical sources

- [README](https://github.com/leemysw/yovoice/blob/main/README.md): Current product overview and supported model families.
- [Model selection](https://github.com/leemysw/yovoice/blob/main/docs/models.md): File sizes, hardware requirements, inputs, voice IDs, advanced options, and license notes.
- [Score guide](https://github.com/leemysw/yovoice/blob/main/docs/score.md): Sound fonts, rendering, MIDI, stems, and score format.
- [Model catalog](https://github.com/leemysw/yovoice/blob/main/internal/catalog/catalog.json): Sizes, hashes, precisions, and download metadata.
- [Source repository](https://github.com/leemysw/yovoice): Implementation and release history.