Hello sorry for the late reply. You could try to send a request to the model worker when the servers are running, using any software (postman for instance). If it replies there, then it is likely a bug in the game I'd be happy to dig into with you.
This is the request Postman request - Pastebin.com
Looks like the request isn't coming back. I guess my processor/GPU just doesn't quite have what it takes. That does make sense, I tend to struggle with heavily quantized llama 7b and usually have to mess with settings to get it to work. Would be cool to see some videos of your game in action on this page though. I really feel like games like this are the future.