I switched to running mistral:7b-instruct locally on my computer because the default endpoint would run out of memory after a few prompts on some of the worlds I'm attempting to create. With this model I noticed the responses beeing even shorter if I use Single-Paragraph event and unless I write fairly short prompts it seems to cue things for upcomming responses and not get through the entire prompt so I turned the settings off which is why I started trying to get the TTS to work. It became alot of reading now :) The model seems to load fine from what I can tell. It's voice selection that seems to not work (unless the model doesn't load properly but no error shows I guess).
1. Initially it'd just keep generating and not finish after selecting a voice and pressing start.
2. Later (I don't think I've changed anything in particular) it started finishing with the generating step fairly quickly and the play button etc shows up. But everytime the length of the wav is 0 and if I play it I get a weird noise for about 1 second and so then it stops.
Other than that I really like the game and look forward to seeing how it progresses.
I have one suggestion though. In the overview you have the World Description. For entities and locations the In-game description. As I understand these they're descriptions meant for the player not the AI.
But traits don't have something similar, one thing I'm doing is using traits for the player to choose who to play which means that sometimes the description becomes rather long it would be nice to have a description field used for when choosing traits where I can sumarize and one for the AI where I can do all the detailing the player doesn't need to see when choosing.