Discord bot for oobabooga webui

This is my first project ever. Any suggestions or contributions are welcome!

8/11/2023 - Chat history is now encrypted. ChatCommands now start with / instead of !.

8/7/2023 - Bots can now produce discord messages longer than 2000 characters by splitting the response into multiple messages

8/4/2023 - Added a /remove <string> command to remove a string from the bot's memory.

1. git clone https://github.com/AICatgirls/aichatgirls.git
2. In the aichatgirls folder (cd aichatgirls), run pip install -r .\requirements.txt
3. You will need to create a discord bot
4. Create a file named .env in notepad and write one line with your bot's discord token in the format TOKEN='your-token-here'
5. Install oobabooga's text-generation-webui. Make sure to add --api to the CMD_FLAGS.txt file.

Text-generation-webui should automatically set the template for your model when you load it. I use TheBloke/Wizard-Vicuna-13B-Uncensored-SuperHOT-8K-fp16 though I'm sure there are better models available these days.

Right now you have to rename your TavernAI character card to character.webp and place it into your aichatgirls folder. If you're using character cards I would love to have your feedback!

• Start Oobabooga's text-generation-webui (start-windows.bat, start-linux.sh, start_macos.sh, or start_wsl.bat depending on your OS)
• Start aichatgirls python aichatgirls.py

Messages you send to the discord bot will now be sent to oobabooga's text-generation-webui and answers returned. To use the bot in a chat channel make sure it has permission and type /whitelist.

You can type /help in the discord chat to get a list of bot commands, and /set to get a list of settings.

Users can now alter bot settings. Usage /set [setting] [value] or /set [setting] to reset a setting to default

• max_response_length (max_new_tokens) - How long in tokens the response can be. Default 400
• min_length - How long you want the responses to be. Default 0
• temperature - A number between 0.1 and 1.9, default 0.4. The higher the number the more creative the response.
• repetition_penalty - A number between 0.1 and 1.9, default 1.18.
• prefix - A hidden phrase that the bot will silently say before giving a response. Default none.