Deploying Bolt.diy with Cloudflare Pages (the easy way!)

Did you fork it to your own repository? And you have to setup Cloudflare to get access to your GitHub account. I believe this was mentioned in the instructions. You will need to setup GitHub integration to access your account.

You may have deleted or transferred the repository associated with this Cloudflare Pages project. For a deleted repository, you will need to create a new Cloudflare Pages project with a repository that has not been deleted. For a transferred repository, you can either transfer the repository back to the original Git account or you will need to create a new Cloudflare Pages project with the transferred repository.

The repository cannot be accessed, this may cause deployments to fail.

You may have excluded this repository from your installation’s repository access settings. Go to your GitHub installation settings:

  • https://github.com/settings/installations for individual accounts
  • https://github.com/organizations/<YOUR_ORGANIZATION_NAME>/settings/installations for organizational accounts

Click Configure on the Cloudflare Pages application. Under Repository access, ensure that the repository associated with your Cloudflare Pages project is included in the list.

See Source: Git integration · Cloudflare Pages docs

Yes i followed each instruction on the manual and video and updates, step by step. i’ll give it a try, thank you.

@rob1 which video do you mean?

Having it installed on my Cloudflare - how to update the code to the latest available on Github?

Take a look at the “Updating Bolt.diy Deployment on Change” section of the instructions.

1 Like

Hello, Thanks For Figuring this whole thing out. I wanted to know, when im finished , how can i deploy?. Boltnew has deploy button but i dont see any here. also whats the point of hugging face or how to utilize that, i saw people downloading deepseek v3 from there, whats the purpose of that. thanks again

There is no automatic deployment option in Bolt.diy, you would need to download the project and deploy it yourself.

Or technically, the same way committing changes to the GitHub Fork would trigger a build and re-deploy automatically, you could do that with your own project… with the push to GitHub option.

But it would require some setup. Though that would be a great workflow for updating changes.

It worked well. thanks a lot.

1 Like


what is the problem why the bolt.diy show error

Have you added a payment method and credits? OpenAI, Deepseek, Anthropic, etc. all requires a payment method. Basically I would just confirm that the API returns a response directly before trying to troubleshoot it as a Bolt.diy specific issue.

Also, I use DeepSeek-V3 through Openrouter personally. It’s the same price, I don’t have to manage as many accounts, they are reliable, they have many models available in one place, give free credits to start, and they offer several Free models to use.

I have same error when I deployed

I followed the instructions, found some LLMs that seems to be working and started building a feature and ran into these messages very quickly. Looks like upgrading to $5/m might solve the problem.

    <h2 class="cf-subheadline" data-translate="error_desc">Worker exceeded resource limits</h2>

What is this $5 subscription fee multiple people have mentioned. No one has answered. Cloudflare has no such subscription.

image

I installed bolt on CloudFlare using these instructions: Deploying Bolt.diy with Cloudflare Pages (the easy way!)

As soon as I started using bolt on CloudFlare I started getting errors in the console about usage limits. I read through the comments of the link above and saw someone signed up for the $5/mo and it seems to fix the issues so I thought I’d try it.

For clarity, I’m throwing stuff at the wall at this point and seeing what sticks as I’ve tried almost everything I can think of.

Cloudflare Pages is free and going to a $5 plan won’t do anything. Plus I didn’t believe they have a $5 plan, hence why I was confused. If you are running into usage limits, likely due to the API you are using. Just sign up for a free HuggingFace account.

Can you explain what the actual issue is? Is Bolt.diy running? Is it an issue with Bolt.diy outputting anything? Because if not, it’s likely the model you are trying to use.

This is the issue: LLMs not working with 31e03ce(v0.0.5) - nightly - #22 by leex279