This repository contains an extension for Oobabooga's Text Generation WebUI that integrates with VTube Studio.
This extension requires the following packages to be installed:
websockets
- Clone this repo
git clone https://github.com/Fire-Input/vtube_studioto theextensionsfolder in the WebUI. - Install the requirements with
pip install -r requirements.txt - Enable the VTube Studio API in VTube Studio's settings.
- Launch the WebUI.
- Start the Text Generation WebUI and enable the VTube Studio extension.
- Enter the port number, and optionally the host of the VTube Studio's API.
- Click the connect button to connect to VTube Studio.
- Allow the extension from the popup on VTube Studio.
- Choose the model, animation, and expression from the dropdown menus.
- If you are running the WebUI with WSL and VTube Studio on Windows, then you should either leave the host blank or you can enter the ip/hostname that WSL uses to point to your Windows device.
- If using WSL, you may need to allow WSL through your local firewall, you can use this PowerShell command on Windows:
New-NetFirewallRule -DisplayName "WSL" -Direction Inbound -InterfaceAlias "vEthernet (WSL)" -Action Allow
This extension is currently a work in progress and may be subject to bugs and other issues. Here are some future plans for the extension:
- Add support for controlling items in VTube Studio.
- Improve error handling and error messages
- Allow for better configuration of the extension
- Add more features for controlling VTube Studio, such as model movement and physics.
- Implement a feature to detect the emotion of generated text and play a corresponding hotkey for more realistic expressions.
- Investigate the possibility of syncing mouth movements with TTS.
- Implement random values for certain parameters like speed or position to add variety to animations.
If you encounter any issues while using this extension, please create a GitHub issue.
Please feel free to suggest any further improvements or features that you would like to see in this extension.