I’m currently working on a web app, and everything was working fine, but now when I ask it to work on a change, it comes back with instructions on how to go ahead & implement that change, its says, “let’s update the CSS to fix the issue,” but then nothing actually happens on the right-hand side…
I tried playing around with the settings, uninstalling & re-installing but the issue persists…
Not sure why this is happening or how to troubleshoot
Tried to export the project and import again? (so you get a new chat history)
I tried to do so but now I’m getting this error again & these are the current settings I have enabled.
I exported the chat & imported it again, but now I can’t even run any commands eventhough I’m using context optimization. I have already uninstalled & re-installed bolt.diy several times just in the past few hours & this problem isn’t going away.
Can’t even run simple commands anymore…
I tried restarting the server as well with no luck.
If I’m able to get a response, its only on the left-hand side & nothing happens on the right anymore.
The error says all. You reached the rate limit of openai, so they block you for further requestes.
maybe consider switch to another Provider and use another Model, or use an provider which also has GPT-4o as Model (OpenRouter, Github
Like I said before, I’m using the context optimization feature as I showed in one of my screenshots. That’s not the issue.
I’m no longer getting that error on different chats.
Now, the same problem occurs on older chats. I get the updated code on the left-hand side but nothing happens on the right-hand side
As said, this is for sure the problem. your chat history is also with context optimization to big as it seems and comes over the limit of chatgpts ratelimit.
The error is clear on this and this errors comes also directly from ChatGPT and not from bolt.diy.
Okay, so what is the best provider I can use so that I dont have to deal with this issue because so far I have like 15+ chats all just trying to work on the same project, but I keep running into this roadblock so I have to start over…
I hear everyone talking about ollama but I’ve never used it. And you mentioned openrouter?
Can I just use openrouter with my own keys or do they provide separate keys for me?
Thanks for clarifying what I was previously confused on & responding so quickly 
Please dont use Ollama. This makes absolutely no sense hosting models on your own hardware for bolt, as they are way to bad compared to cloud models.
OpenRouter => you need to by credits there and do seperate API keys from their page
I personally use Google Gemini 2.0 a lot cause it has very huge context size with 2M token and its free. Didnt reach a limit there yet.
You can also take a look at some of my videos on youtube if you didnt do: https://www.youtube.com/channel/UC_a85mUHqsy5j0CYCgLnkEQ
In the Installtion video of Windows, I also show how to get the API key for google and how to use.
Hey again, took everything you said into consideration & now I’m making excellent progress using Google Gemini 2.0 Flash like you recommened. I also subscribed to your YouTube channel & saw a few of your videos including the most recent one which basically showed all the issues I was running into with Deepseek — too many errors with the preview page exactly how you experienced.
That’s why I decided to use Gemini instead & now I’m also using CodePen.io to help me provide examples to Bolt.diy on exactly the look & style I’ going for. I still experience a few bugs here & there but its much more manageable & no real issues with the previews
Thanks again for all the guidance. I’m grateful to come across this line of support during my 1st weekend playing around with all of this as a complete beginner. I admit that this can be very overwhelming at first, but if you just keep persisting & testing things out you learn along the way & it becomes a lot easier than you expected. Plus, the support here has really helped me finally get things rolling on my designs.
… now I just need to learn how I can input them into platforms like GHL or Shopify with full functionality.
I’m also using Make.com automations & would like to be able to integrate them seamlessly… or perhaps even create native functionality so I don’t have to use additional platforms like Make for the automations infrastructure.
thanks for this amazing post and happy it helped you out.
Is there a question I should answer. I am not sure 
