Generate speech from text
Use a compatible GPT-SoVITS setup to turn a script into speech for narration, dubbing, dialogue, or other audio projects.
Browse GPT-SoVITS models that can generate speech from text in a compatible local setup. Download each model from the source where its uploader shared it.
Browse GPT-SoVITS voice models →Showing 24 of 148 · Page 1 of 7
Nathaniel Bandy GPT-20 Model-15
by coolwowsocoolwow
by coolwowsocoolwow
Mario GPT SoVITS
by coolwowsocoolwow
by jamessjunk
Oblina - GPT-SoVITS,
by xhepyxopila
by coolwowsocoolwow
Protegent Ad Voice 2017 GPT SoVITS
by coolwowsocoolwow
by jamessjunk
Franklin Clinton
by bruh110
by coolwowsocoolwow
Cave Johnson - JK Simmons GPT-SoVITS
by lusbert
by kakao111
Genshin Impact · Japanese-language version
by kakao111
Peppa Pig GPT SoVITS
by coolwowsocoolwow
Brooklyn T Guy
by coolwowsocoolwow
TF2 Spy Dennis Bateman_e16-GPT_e15-SoVITS
by lusbert
Troy Mcclure
by loren85
English-language version
by kit-lemonfoot
Chainsaw Man
by loren85
GLaDOS Portal 2 -GPT_e20-SoVITS
by lusbert
by kakao111
English-language version
by kit-lemonfoot
Duke Nukem GPT-SoVITS, 20-
by jamessjunk
by loren85
Continue by voice, framework, language, franchise, category, or creator when that context is represented in these models.
These labels come from models currently shown in this directory. Use them to distinguish a character's game, series, dub, or other recorded version before choosing a download.
Start with what you want to make: live voice changes, recorded speech, a song cover, or another vocal project. Then check the model's language, version, files, and app requirements.
Different versions of the same framework may need different tools or settings.
Do not assume a model's language from the character name alone.
Look for the required weights, configs, checkpoints, setup notes, demos, and app requirements.
No. Each listing links to the external page where the model was published.
Different creators, datasets, framework versions, and training runs remain separate models.
GPT-SoVITS is used for text-to-speech and voice cloning workflows. A model may include separate GPT and SoVITS checkpoints, and results depend on the software version, reference audio, language, and generation settings.
AIVoices checks that each listing points to a live model source and looks for the paired checkpoint files used by GPT-SoVITS. Read the original model card for installation steps, supported languages, and license terms.
Check the intended language, checkpoint version, reference-audio requirements, and source instructions before downloading.
Use a compatible GPT-SoVITS setup to turn a script into speech for narration, dubbing, dialogue, or other audio projects.
GPT-SoVITS packages commonly include both GPT and SoVITS checkpoints. The source instructions should explain which files work together.
Do not assume every model supports every language. Use the uploader's model card and any available samples as your guide.