InvalidContentLength on PDF without any reason

Question

InvalidContentLength on PDF without any reason

M. Kerr 0

I'm trying to set up a basic pipeline to take a PDF, OCR it, and return a searchable PDF with Azure prebuilt-read under the v4.0 API. I'm sure ten thousand people have done this before, right?!

I can't get Azure to process anything more than the smallest PDFs. For example, I have a 12.1 MiB PDF of business documents (US Letter size; about 100 pages; no more than 5000x5000 pixels per page) that Azure DI refuses to read. "InvalidContentLength - The input image is too large. Refer to documentation for the maximum file size." I am on S0; my documents are far under the maximum size.

First, I tried submitting files with REST. It got above 10 MB or so and complained that the file was too big. So the Internet told me that I had to set up an unneeded Blob Storage instance and pass a URL from there. OK, fine, so I did that and started submitting a urlsource instead. Same error.

Even in the online Document Intelligence Studio OCR/Read, my document uploads fine (or accepts my Blob Storage URL), even shows a correct preview window where I can scroll through the document, but I get the exact same !!@#$!@ error when I "Run analysis." Even if I ask it only to analyze one page. "InvalidContentLength"

I am on the paid version which is supposed to support 500 MB PDFs.

I am so, so incredibly frustrated. This error message is not helpful! The documentation says it should work fine!

I would think that there would be an easy template or complete how-to document for something so simple.

Deleted

This comment has been deleted due to a violation of our Code of Conduct. The comment was manually reported or identified through automated detection before action was taken. Please refer to our Code of Conduct for more information.
Manas Mohanty 5,350 Reputation points Microsoft External Staff Moderator

2025-06-12T17:43:41.6066667+00:00

Hi M. Kerr

Could you please let us know if new document intelligence in other regions fixes the issue for you.

Requesting few more pdfs to validate our observation

Thank you.
Manas Mohanty 5,350 Reputation points Microsoft External Staff Moderator

2025-06-13T17:24:04.9466667+00:00

Hi M. Kerr

We were not able to hear from you. Hope the pointers shared helped address your issue.

Thank you.
M. Kerr 0 Reputation points

2025-06-16T11:59:18.22+00:00

Hello,

As I mentioned, there is an open support request that I believe has been passed to back-end engineers. I am waiting for their response.

I did create a resource directly as S0 in another region and it does seem to work properly, but they also confirmed that the issue is real with my original resource.

Thank you.
Manas Mohanty 5,350 Reputation points Microsoft External Staff Moderator

2025-06-16T17:18:02.41+00:00

M. Kerr

Glad to hear that you were able to overcome the issue in a new resource.

Could you accept the answer from me on deploying in other regions.

I think that particular resource is corrupt on backend and will be analyzed for RCA by product team.

Thank you.

2 answers

Your answer

Deleted

This comment has been deleted due to a violation of our Code of Conduct. The comment was manually reported or identified through automated detection before action was taken. Please refer to our Code of Conduct for more information.
Manas Mohanty 5,350 Reputation points Microsoft External Staff Moderator

2025-06-12T17:43:41.6066667+00:00

Hi M. Kerr

Could you please let us know if new document intelligence in other regions fixes the issue for you.

Requesting few more pdfs to validate our observation

Thank you.
Manas Mohanty 5,350 Reputation points Microsoft External Staff Moderator

2025-06-13T17:24:04.9466667+00:00

Hi M. Kerr

We were not able to hear from you. Hope the pointers shared helped address your issue.

Thank you.
M. Kerr 0 Reputation points

2025-06-16T11:59:18.22+00:00

Hello,

As I mentioned, there is an open support request that I believe has been passed to back-end engineers. I am waiting for their response.

I did create a resource directly as S0 in another region and it does seem to work properly, but they also confirmed that the issue is real with my original resource.

Thank you.
Manas Mohanty 5,350 Reputation points Microsoft External Staff Moderator

2025-06-16T17:18:02.41+00:00

M. Kerr

Glad to hear that you were able to overcome the issue in a new resource.

Could you accept the answer from me on deploying in other regions.

I think that particular resource is corrupt on backend and will be analyzed for RCA by product team.

Thank you.

Answer 1

Manas Mohanty 5,350 Microsoft External Staff Moderator

Hi M. Kerr

I tested in Central US region with the 3-page pdf shared from you.

Issue seems to be specific at your resource side.

User's image

Could you create a new document intelligence resource (with no underscore or special character in the name) in any other supported region and let us know if the issue persists.

Status at customer side - Resolved with new deployment with another region.

Thank you.

Answer 2

Hi M. Kerr

Thanks for using the Q&A platform.

The error InvalidContentLength suggests the input image is too large based on https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/how-to-guides/resolve-errors?view=doc-intel-4.0.0

However, the error is commonly seen by users even when working with smaller files if you have special characters in filenames, or a corrupt PDF structure, or unsupported features embedded in the file.

Kindly validate your PDF using PyPDF2 or PyMuPDF to ensure it’s not corrupt, has pages, and isn't encrypted. Verify page image dimensions by extracting each page image via pdfplumber, pymupdf, or poppler, and check that none exceed 10,000 pixels in width/height.

Find a similar issue stackoverflow: https://stackoverflow.com/questions/79272102/azure-document-intelligence-formrecognizer-invalidcontent

If the response was helpful, please feel free to mark it as “Accepted Answer” and consider giving it an upvote. This helps others in the community as well.

Regards,

Obinna.

M. Kerr 0 Reputation points

2025-06-09T10:06:38.5566667+00:00

Hello,

I did verify that all pages were < 5000x5000 and none of the PDF tools I've used on it have given any errors. It's also happening with many PDFs; not one specific item.

The Stack Overflow link is for a "The file is corrupted or format is unsupported" message. That's not the error I'm getting.

At the moment, my hypothesis is that my DI resource is still being treated as a free F0 status with a 4 MB limit even though it's S0. I'm waiting for support to confirm this.

Thank you

Share via

InvalidContentLength on PDF without any reason

2 answers

Your answer