it’s literally ridiculous
Also, can I just remind everyone that we pay for this service, yet there always seem to be ill-thought-out mistakes made constantly.
3 Likes
I have created a simple request forwarder that you can implement quickly and use.
To use it is very simple. Add a header to your API call called destinationURL.
This should be the API call URL(e.g https://api.openai.com/v1/images/generations).
Then, change the API call URL to https://us-central1-not-quite-unicorns.cloudfunctions.net/forward_request.
Your Authorization header remains the same and can be a header or shared header.
It will return the body of the API you’re calling, though not the headers. Don’t re-initialise the API call. When the issue is fixed, you can just copy the destinationURL header into the API call URL and get rid of the destinationURL header.
This is published on a no guarantees basis so use it if you wish or don’t.
5 Likes
wow that was fast! thanks for making this! i’ll test it out and let u know if it all works for me!
1 Like
so i shouldn’t re-initialize the call? what will happen if i do?
1 Like
It’ll just create a pain in the ass
ahahahahahahahaha
looks like my API calls just started working again, so I’ll hold off on implementing this fix for now
1 Like
This is gold. It happens each and every time lmao. React with the bad emoji and tell them this.
1 Like
This issue has been resolved (for now). We made a good effort in contacting Bubble and OpenAI and it seems one of the parties have made some changes, though I do not have confirmation on this, the API calls for DALLe is now working.
Happy building everyone!
Update: Not fully resolved. Still looking for an official solution from either parties. I suppose API calls ATM is just be low so we get/got a response.
2 Likes
thank the lord 
1 Like
although, this did happen to me yesterday as well. then it started working again.
then today it was broken once again, and just now restarted working.
so it could be intermittent? you said we don’t have confirmation that bubble and openai talked right? so this could just happen again
Yes, I unfortunately still get the error when troubleshooting further, it’s a 50/50 if it goes through. Still work to be done but better than a 100% fail-rate. I suppose it is just lots of Bubble apps using the OpenAI DALLe API and Bubble being rate-limited – calls now might just be low so we get a response, hopefully not though. Still following and awaiting an official response from the Bubble team.
1 Like
I went through a progression of deteriorating emoji responses with a corresponding increase in attitude. 
1 Like
I think that means that now we’re on TWO whole servers.
1 Like
This will work, guys! Here’s the source code if you want to self host (which I’d recommend so you can always fall back on it if the issue reoccurs…
You will continue to intermittently experience this issue if you’re not using client-side calls until OpenAI or Bubble changes something, which likely won’t happen any time soon. A simple request forwarder will fix the issue for you 
1 Like
Could it possibly be related to the User-Agent?
1 Like
This topic was automatically closed after 14 days. New replies are no longer allowed.