Skip to main content

Live test - Max retry async calls reached

Feature Request

I’m trying to configure a model to process tables of data from multi-page PDFs. The table structure isn’t uniform as some of the header titles have a different orientation (set vertically).

The error message that returns is "Max retry async calls reached".

Even when I reduce the PDF to a single page PDF it still does not return an output.

Is there any particular way I should be configuring my model for a successful extraction?

Posted by Damien St. John ·
Status: Completed4 comments

Log in to comment and vote

Comments4

  • Maxime Churin

    Team•

    Aug 14, 2025

    Hello, we just release the new UI for the live test. In this new version, even if live test fails due to polling delay, you have a new tab documents history where you can retrieve the result after processing is done.

    • Damien St. John

      •

      Aug 14, 2025

      Sweet! Thanks for the update. It’s working perfectly in my project so far 🙌🏽

  • Ianaré Sévi

    Team•

    Aug 1, 2025

    As mentioned by @Jonathan Grandperrin , currently your best bet is to use the API via one of our client libraries rather than the Live Test.

    You can make changes to the Polling Options as needed to avoid “max retry” errors. The documentation is here:

    https://docs.mindee.com/integrations/client-libraries-sdk/configure-the-client#polling-configuration

  • Jonathan Grandperrin

    Team•

    Aug 1, 2025

    Hi there! If you are seeing this message when uploading the document in the live test, it’s because on the live test UI, we have set a time processing limit. Your data schema seems pretty long so maybe the processing time hits this limit.

    Thank you for pointing this out, we are actually already working on a better way in the live test to upload the documents, making sure no retry limit can be reached so that you can see your document extraction. It will look like the following screenshot (all the documents sent to the live test will be processed, and retrievable asynchronously in the documents history tab)

    That being said, If you want to get the extracted data, you can still use the API. You’ll be able to increase the MAX_RETRY constraint or even use a webhook to get your results.

    Best regards