Text‐generation‐webui Remote Extension

Aside from installing AllTalk as part of Text-generation-webui, AllTalk can be integrated with Text-generation-webui as a remote extension, allowing you to use AllTalk's TTS capabilities without installing it directly within Text-generation-webui's Python environment. This setup is ideal for users who want to run AllTalk on separate hardware or in its own Python environment.

Features

• Connect to a remote AllTalk server from Text-generation-webui
• Access all AllTalk TTS features through Text-generation-webui interface
• Maintain separate environments or hardware for Text-generation-webui and AllTalk

Installation

After you have installed a Standalone AllTalk server

1. Locate your Text-generation-webui installation directory.
2. In the Text-generation-webui extensions folder, create a new folder named alltalk_tts.
3. Copy all files from within \alltalk_tts\system\TGWUI Extension\ folder into the newly created alltalk_tts folder.

text-generation-webui/
      ├── extensions/
      │   └── alltalk_tts/           # Create this folder
      │       ├── script.py          # Copied from \alltalk_tts\system\TGWUI Extension
      │       ├── settings.py        # Copied from \alltalk_tts\system\TGWUI Extension
      │       └── [other files]      # Copy all other files from \alltalk_tts\system\TGWUI Extension
      ├── models/
      ├── characters/
      └── [other text-gen-webui folders]

Setup

1. Start or restart Text-generation-webui.
2. In the Text-generation-webui interface, find and select the "alltalk_tts" extension in the list of available extensions.
3. Save your settings and restart Text-generation-webui.
4. Upon restart, the AllTalk extension will load. You may see an error or warning if the IP address of your remote AllTalk server is not yet configured.

Configuration

1. Option A In the extension settings, enter the correct IP address and port of your remote AllTalk server's API address. If its a local machine you can use 127.0.0.1:{port-number}, if its on a specific IP address, enter that IP address and port number e.g. 192.168.1.240:{port-numner}.
2. Option B If your AllTalk server is accessed through a secure tunnel on the internet, use https:// instead of http:// in the protocol dropdown and enter the tunnel URL address.
3. Click the "Refresh settings & voices" button to download all available settings from the AllTalk server.

💡 Tip: The API port number is shown on the console splash screen and the default is 7851.

Usage

• The first successful TTS generation will automatically save your settings to a local configuration file.
• Use the "Refresh settings & voices" button to ensure the extension is in sync with your AllTalk server, especially after making changes on the server.
• If certain settings or features are unavailable (greyed out) in the extension, it may be because the loaded TTS Engine that's running on the AllTalk server doesn't support them.

Troubleshooting

• If you encounter connection issues, verify the IP address, port, and protocol settings.
• Check firewall settings to ensure Text-generation-webui can communicate with the AllTalk server.
• Review the Text-generation-webui console for any error messages related to the AllTalk extension.