Sometimes, quite often actually the AI stops creating files, and sometimes it cuts off in the middle of generation. Not quite sure if its model problem, I’m using Google Flash 2.0 thinking experimental.
I guess you are on the stable branch?
What branch is that? because i dont really trust what my command prompt says, although it says main. Is that correct version?
thats the latest main, yes.
Make sure you enabled the settings as shown here:
It was enabled. Btw, any chance for you to see how to make Bolt work well with GitHub API? I have a feeling that could be the problem why its not working well. Because sometimes it works well, sometimes it doesn’t. I can give you paid access token for testing if you would like. My local server does support chat completions and /models. it doesnt support /v1/models or all those 404 api requests. Are they important?
I dont know what exactly you mean. Can you show your full setup (provider, chat history screenshot, model you are using, …)?