Using Groq and Llama 3.1B
PC running Node and not running Docker as yet…
Bolt.diy won’t automatically run any of the build and npm starts…
Only a new installation here and working out the kinks but having a few issues including finding the working folder for this code.
Trying to use similar commands from another guide like ‘npm insall pnpm’ via Bolt terminal but not getting results.
I have API keys with Groq, OpenRouter and just want to see something run here like Bolt.new
Please help.
Hi @shawn,
try bigger Models, not such small with 8B :). They dont provide that good code or you have to let them fix very often until they get it right.
Use e.g. Qwen 2.5 Coder Instruct 32B from OpenRouter.
OK. This works but seems the only one I can get working.
I can’t seem to get any of the other options working on this. If that’s the case, why are they there. I’ve got this to start working. Will see how we go.
I was suggested OpenRouter and Gemini as Gemini is free but can’t get any of those to work with OpenRouter and my API.
Thanks.
Bolt just lists what the providers provide as models. This does not mean that they work well mit bolt. There is no filter etc. You are free to choose from everything your configured provider has.
So you have to choose the right model for the right task 
Some suggestions mentioned in other topics are:
I use Gemini Flash 2.0 (from google directly) at the moment for my current project and it works well
@leex279 cheers Thomas.
I need to trawl thru other users projects and settings to get a feel for where things are too.
Will be interesting to see where this goers in the coming weeks.
I hope to be active here and help others once I’ve cleared out these first few teething problems. Even thought of building a page to deploy locally that can do health checks and identify components working and clickable fixes. 
I spent the entire time using it, and I’m always just getting no preview available as well. So I don’t know if this means Qwen 2.5 Coder is the only one that works. I haven’t tried that, but I’ve been using Claude. And I figured that is going to be a bigger model, but running into the same issue.
Which provider you using…?
Do you have an API key and making connections…?
It’s making connections and its costing me API credits. This is for Claude. Except I don’t see anything for the preview.
Maybe clear your chat history and use the optimized prompt option in the Bolt.diy settings.
I checked off the optimized settings box.
I also removed every single chat I had prior.
It seemed to have done the trick in removing the no preview. I actually have something that has popped up.
Thank you!
Weird behavior I know, but I ran into the same thing. Should probably note it and put it in a knowledge base, troubleshooting guide, or something. Thanks!
