This site does not pay for the AI that writes worlds: use a model that runs in your browser (free, once downloaded), or your own key (an OpenRouter account, or your own OpenAI-compatible server).
Any OpenAI-compatible server works, a local one too, as long as it lets this page call it (CORS): llama.cpp's llama-server does already; Ollama needs OLLAMA_ORIGINS=*; LM Studio: turn on “Enable CORS” in its server settings. Your browser may ask to reach your local network.
llama-server
OLLAMA_ORIGINS=*
Kept in this browser only, and sent only to the provider you chose: never to this site's server.