Run your own AI chat bot locally. 100% private. 100% yours. Follow these steps:
Clone or download the web UI from the official repo:
Oobabooga GitHub
Follow the README
to install Python and required packages (like torch
).
We recommend: writing-roleplay-20k-context-nemo-12b-v1.0-gguf
Available here: Hugging Face
Q4_K_M
or Q5
depending on your GPU and RAM capacity.
Move the downloaded .gguf
file to:
text-generation-webui/models/
You can create a folder inside for organization if needed.
Open a terminal in the Oobabooga folder and run:
python server.py --model your-folder-name
Wait for it to load, then open http://localhost:7860
in your browser.
You’re all set. The bot runs completely offline with full erotic roleplay support.
No filters. No cloud. No limits.
Try adding features like long term memory, long-form generation, or character cards using Oobabooga extensions.
See the Extensions tab in the UI for options.
(Guide coming soon)
If something breaks, ask for help in the GitHub Discussions tab or check back here soon — we’re building a troubleshooting section.
Or just grab the $5 one-click installer and be done.