Azure Form Recognizer to not read certain pages from PDF

Question

Azure Form Recognizer to not read certain pages from PDF

E-R 0

Correct me if I'm wrong but if you give PDF to parse for Azure Form Recognizer then it will try to parse all the pages and parse the information it has been trained to parse. And the price is determined on how many pages the model will read through.

So this means that even if the PDF has only 1 out of 8 pages that has the information it has been trained to parse it will still read the other pages. And there for the costs will be for 8 and not 1 pages. Is this correct?

If so is there any way to configure the model to only read pages that have something to parse?

I am building a Power Automate flow that uses Analyze Document for Prebuilt or Custom models (v3.x API) -connector. That has Pages parameter where I could set up certain amount of pages. But since I don't always know what pages are relevant I can really use that. User's image

dupammi 8,615 Reputation points Microsoft External Staff

2024-08-09T09:28:36.9933333+00:00

Hi @E-R

Thank you for your question.

Yes, Azure Form Recognizer will attempt to parse all pages in a PDF by default, which means that the cost will be calculated based on the total number of pages processed. However, you can reduce costs by using the "Pages" parameter to specify which pages to analyze, but this requires you to know the relevant pages beforehand. If you can't determine which pages are relevant in advance, the model will process all pages, leading to higher costs.

For more details please refer - FAQ#can-document-processing-handle-multiple-forms-in-a-single-document and page-range
dupammi 8,615 Reputation points Microsoft External Staff

2024-08-10T00:50:58.5033333+00:00

Hi @E-R

We haven’t heard from you on the last response and was just checking back to see if you got a chance to check my above suggestions. Thank you.
E-R 0 Reputation points

2024-08-12T13:33:08.0433333+00:00

Hi @dupammi

Thank you for your answer that confirms my suspicions. So indeed there is no way to reduce the cost unless you know before hand what pages you want to parse.
I was hoping there was some configuration for example to first parse the middle page (so if there is 10 pages it would start from 5) and then if that doesn't have anything to parse it would not parse anything after that but only the first half (so first 5 pages).
Or better yet there would be configuration to make it stop parsing after the first page that didn't have anything to parse.

1 answer

Your answer

dupammi 8,615 Reputation points Microsoft External Staff

2024-08-09T09:28:36.9933333+00:00

Hi @E-R

Thank you for your question.

Yes, Azure Form Recognizer will attempt to parse all pages in a PDF by default, which means that the cost will be calculated based on the total number of pages processed. However, you can reduce costs by using the "Pages" parameter to specify which pages to analyze, but this requires you to know the relevant pages beforehand. If you can't determine which pages are relevant in advance, the model will process all pages, leading to higher costs.

For more details please refer - FAQ#can-document-processing-handle-multiple-forms-in-a-single-document and page-range
dupammi 8,615 Reputation points Microsoft External Staff

2024-08-10T00:50:58.5033333+00:00

Hi @E-R

We haven’t heard from you on the last response and was just checking back to see if you got a chance to check my above suggestions. Thank you.
E-R 0 Reputation points

2024-08-12T13:33:08.0433333+00:00

Hi @dupammi

Thank you for your answer that confirms my suspicions. So indeed there is no way to reduce the cost unless you know before hand what pages you want to parse.
I was hoping there was some configuration for example to first parse the middle page (so if there is 10 pages it would start from 5) and then if that doesn't have anything to parse it would not parse anything after that but only the first half (so first 5 pages).
Or better yet there would be configuration to make it stop parsing after the first page that didn't have anything to parse.

Answer 1

Hi @E-R

Thank you for your follow-up and for confirming your understanding. You're correct that currently, Azure Form Recognizer doesn't offer a built-in configuration to automatically skip or stop processing pages based on the content found during parsing. The "Pages" parameter requires pre-specifying which pages to analyze, so it can't dynamically adjust based on the content of the pages.

I appreciate your suggestion about starting from a specific page or stopping after finding an irrelevant page—this could be a valuable feature for optimizing costs and efficiency. I encourage you to submit a feature request through the Azure Feedback Forum. Here's the link to the Azure Feedback Forum: Post idea · Community (azure.com).

Thank you again for your time and patience throughout this issue.

Please don’t forget to Accept Answer and Yes for "was this answer helpful" wherever the information provided helps you, this can be beneficial to other community members.

Joseph Premanand.M 0 Reputation points

2025-12-11T07:57:17.4033333+00:00

HI Microsoft Team,

Could you please confirm, Azure AI Document Intelligence pre-built invoice model have capability to identify the document types as Invoice, Delivery note, PO and supporting documents? And instead of single whole document to OCR and identify document type, can it be possible to set page range to extract 1-15 pages out of 100 pages in single document.

Regards,

Anand

Share via

Azure Form Recognizer to not read certain pages from PDF

1 answer

Your answer