Custom Extractor Page Limit

[frank90089]

The limits of  a custom extractor  are:

Maximum pages (online/synchronous requests):	15
Maximum pages (batch/offline/asynchronous requests):	200
Maximum pages (imageless mode online/synchronous requests):	30

Now i want to extract a document with 16 pages. I use imageless mode.

 i got a error " Document pages exeed the Limit 15 got 16"

If i use the same Code with a Invoice Parser everything works fine.

[dawnberdan]

Hi @frank90089,

Welcome to Google Cloud Community!

Custom Extractor in imageless mode should support up to 30 pages, but you're encountering a limit of 15 pages instead. This could be due to a configuration issue or a restriction specific to your setup. Here are some things to check and try in order to resolve the issue:

1. Verify Imageless Mode is Enabled:

• Make sure that imageless_mode is explicitly enabled in your ProcessRequest. All processors can now extend the maximum page limit for online and synchronous requests to 30 pages. To do this, activate imageless_mode in the ProcessRequest.
• According to the Document AI release notes, to use this feature with the Custom Extractor, you must first submit a request to be allowlisted by completing the form Allowlist Request for 30 Page Limit in CDE.

2. Check Processor-Specific Limits:

• Different processors have different page limits. The Invoice Parser may have a higher threshold, which is why it works while the Custom Extractor does not. Check this document for more details.

3. Use Batch Processing Instead:

• If synchronous processing is hitting a limit, consider switching to batch processing, which allows up to 200 pages.

In addition, You may refer to this case, as some users have faced similar issues with exceeding page limits in Document AI. It contains insights that could be helpful for you.

Was this helpful? If so, please accept this answer as “Solution”. If you need additional assistance, reply here within 2 business days and I’ll be happy to help.

[frank90089]

Hello,

Thanks for your reply. Regarding the individual points:
1. Imageless Mode is enabled and works without problems for the Invoice Parser. It doesn't work with the Custom Extractor.
2. I checked the processor limits. Documents with more than 15 pages should work asynchronously, but they don't.
3. I tried batch processing, but I get the same result. The limit is 15 pages.

What I also did was fill out and submit the Allowlist Request for 30 Page Limit form in CDE three weeks ago. But I haven't received a response yet.