Inference API not working: Error 429 too many requests

I have a Postman setup to make inferences using the Roboflow API and it was working fine. However the next day I tried again and got a 429 error implying “Too many requests”. I’ve been blocked since and not able to work around this error. Even in the Example Web app uploading an image works and shows inferences but using a URL throws an error.

Hey there @Christian_Rivera

Could you please provide the following so we can help replicate the issue? (so we can help you troubleshoot)

  • Your project’s Universe link or workspace & project ID
  • The endpoint you were requesting to
  • A screenshot or link to the Postman request (make sure to remove or redact your API key)

Hey @leo
Thanks for reaching out. Here’s my project’s info:

Project ID: IPhfhCrXpbfcTRX4h9bVOKDhGAf1

Hey @Christian_Rivera

Thanks for providing the updated info here. There was a user that experienced similar issues here:

The 429 error isn’t with our API, it’s with, which we can’t control. As suggested in the quoted answer above, if you aren’t having issues accessing the image locally, I recommend you query the image data locally, then POST the image data to the Roboflow Hosted API.

Hey @leo,
Using a local image inference was my first approach. I took a picture with the expo app module, convert it to base64 and send a post request to the Roboflow API. However, I always received “Internal error” while doing that. This is why I opt out using the imgur approach but now I have the 429 error.

Do you know what am I doing wrong for local image API call? Is there a Postman example I can follow?

Thanks in advance.

Here’s a postman example of a base64 local image:

And here’s using a local file as it is:

Both methods fails on postman and also doing a react native axios call or fetch
I would appreciate some guidance @leo

Hey @Christian_Rivera

It looks like you are passing the image’s base64 string in as form data. Could you try passing the image string directly, with no attributes and body type? See our object detection hosted API reference.

Hey @leo appreciate very much your help. That did work:

I’ll leave this here so people can check it out in the future. #youRock

Awesome, glad to hear its working!

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.