Graph API - Docx to Pdf Conversion .

Question

Graph API - Docx to Pdf Conversion .

Naresh Namani 0

Hi , we have one word file , we are trying to convert docx to pdf, we seeing pagination issue.

if you see below word.png , page ended with "that is not" line , and next page starts with "caused by the " .

But in PDF version , page ending at "cancellation of any " and next page starts at "export of other " .This is not matching with docx file we have .

word.png pdf.png

0 comments

1 answer

Your answer

Answer 1

Anonymous

Hello Naresh Namani !

This pagination mismatch between DOCX and PDF when using Microsoft Graph API is actually pretty common, and there are a few things that could be causing it.

When you use the Microsoft Graph API to convert DOCX to PDF (via the /convert endpoint), it processes the document through Microsoft's cloud rendering service. The pagination differences you're seeing happen because the Graph API's PDF conversion engine makes its own decisions about page breaks based on the document's content, formatting, and target output settings, which don't always match what you see in Word.

My recommendations:

Check your Graph API call parameters - Make sure you're not passing conflicting page size or margin parameters in your conversion request. The Graph API uses default settings if you don't specify them.
Font embedding issues - If your DOCX uses custom fonts, the Graph API might substitute them during conversion, causing text to reflow differently. Try using standard system fonts or ensure your fonts are properly embedded in the DOCX.
Page setup in the source DOCX - Set explicit page dimensions and margins in your original DOCX file before uploading it to OneDrive/SharePoint. The Graph API conversion respects these settings but might interpret them slightly differently than Word.
Try the Graph API with specific parameters - Use the /convert endpoint with explicit format parameters:
```
   POST /me/drive/items/{item-id}/content?format=pdf
```

Workaround: Since you're already using Graph API, you could try converting to PDF first, then use another Graph API call to convert that PDF back to DOCX if you need both formats to match exactly. The PDF-to-DOCX conversion might give you more predictable results.

Alternative: If exact pagination is critical, consider using the Office 365 APIs directly or the Microsoft Graph Excel/Word APIs with more granular control over the conversion process.

Hope this helps point you in the right direction!

Naresh Namani 0 Reputation points

2025-09-24T06:25:07.77+00:00
Hi Raymond,

sorry if im not wrong ,it is not POST request right. To convert we need to use GET request type and provide format . Followed below Microsoft Graph Guide.
https://learn.microsoft.com/en-us/graph/api/driveitem-get-content-format?view=graph-rest-1.0&tabs=http
GET /me/drive/items/{item-id}/content?format={format}

GET /me/drive/items/fileName/content?format=pdf

First file is getting uploaded to onedrive (PUT ) and then using GET converting to PDF format.

PUT https://graph.microsoft.com/v1.0/me/drive/root:/"+ fileName + ":/content

Once file is uploaded to Onedrive , i have checked how it was in one drive. File is same exact like in local, But after conversion , it is not same as .docx file,getting pagination issue.We want to integrate this conversion process in our Java application. Pagination is critical for our use case , can you suggest some alternatives . Office 365Api's are provides services for OneNote, Outlook,Excel, OneDrive,Teams,Planner and sharepoint and it is exposed to Microsoft Graph. Microsoft Graph Excel/Word APIs are the one im using correct ? if it is not ,can you please provide any documentation link. when we try to convert pdf to docx using , getting 406 Not Acceptable response.

GET /me/drive/items/fileName.pdf/content?format=docx
Anonymous

2025-09-24T08:39:09.85+00:00
Hello Naresh Namani ,sorry if I made mistakes in my earlier answer.

You’re right about the method. The pattern is:

Upload
PUT https://graph.microsoft.com/v1.0/me/drive/root:/path/to/file.docx:/content

Convert to PDF
GET https://graph.microsoft.com/v1.0/me/drive/items/{item-id}/content?format=pdf

Docs (supported formats and expected status codes): Get driveItem content in a specific format

About PDF → DOCX: that direction isn’t supported on this endpoint; a 406 Not Acceptable is expected.

If pagination needs to match Word closely, a few product‑only options that tend to be more faithful than the generic convert endpoint:

Use Word Online (Business) “Convert Word Document to PDF” via Power Automate or Logic Apps. It uses the Word Online engine and often preserves page layout better. You can trigger a flow via HTTP from your Java app, store the PDF in OneDrive/SharePoint, then download it.

Connector docs: Word Online (Business) connector

If you stay with the Graph API conversion, tighten the source DOCX before upload:

Use standard Office fonts (Calibri, Cambria, Arial, Times New Roman) to avoid font substitution.

Remove Compatibility Mode (File → Info → Convert), then re‑save.

Set explicit paper size, margins, and header/footer distances in Page Setup.

Review paragraph pagination controls on critical text: Keep with next, Keep lines together, Widow/Orphan control.

Prefer inline objects over floating/wrapped text boxes when exact pagination matters.

A fully product‑only pipeline without third‑party libraries: Java app → call a Power Automate HTTP‑triggered flow (Word Online (Business) action converts DOCX → PDF) → store in OneDrive/SharePoint → your app downloads the PDF.

Hope this solves your problem.
Naresh Namani 0 Reputation points

2025-09-25T07:33:44.9433333+00:00

Yes i have tested with word online "Convert word Document to pdf " via power automate ,But it is also generating same output like MS Graph API. It is not converting docx to pdf accurately , pagination issue is present.

In our windows application, when we open docx file and we print -> save as PDF , it is converting to pdf in a expected pagination format.
But MS Graph and Online Word Converter is not converting as expected. there is pagination issue .

Question is : Windows MS Word and Online Word uses different Engine for conversion ??

Is there any alternative solution to convert accurately using any api / other api's ?

Please refer below :
Word file :

PDF File :
Anonymous

2025-09-25T10:07:01.6966667+00:00

Hi @Naresh Namani , thanks for reaching on Microsoft Q&A!

Actually, the behavior you are encountering is expected, even if you use something else rather than Graph API. And of course, the issue can usually reproduce when you export PDF using Microsoft Words.

A Word DOCX is flow-based, while PDF is fixed-layout, so rendering engine used by Microsoft Graph can produce different line breaks and pagination compared to what you see in Word. This can usually cause by font, margin handling, or differences in renderer.

If you think my answer is useful, please interact with the system, thank you!
Naresh Namani 0 Reputation points

2025-09-29T16:43:54.29+00:00

Hi ,

Thanks for responding. We are looking for solution to match generated pdf with docx accurately. Above in alternatives, mentioned Office 365 API's , can you provide any documentation on how we can use office 365 api's and how we can convert files into pdf's and integrating into Java application.

We will test with office 365 api's to convert or figure out some alternate possibilities .

Thank you,

Naresh
Anonymous

2025-09-30T07:52:45.1866667+00:00
Hi Naresh Namani ,

Thanks for the follow-up. To clarify:

Microsoft Graph / Word Online / Word Automation Services all rely on the cloud rendering engine. It’s very good for most cases, but it does not always match Word desktop pagination. Differences typically stem from font substitution, line-breaking, and layout nuances in the service renderer. Since you already tested with Word Online (via Power Automate) and saw the same page shift, Word Automation Services would behave similarly and won’t resolve this particular fidelity requirement.

Word desktop application (locally installed) is the only option that consistently guarantees the on-screen pagination you see while editing. If pixel‑perfect pagination is business‑critical, the recommended approach is to automate PDF export through Word desktop (e.g., COM Interop, an Office Add-in running locally, or a controlled workstation/VM process scripted via PowerShell).

For Java integration, if slight pagination differences are acceptable, you can still use Microsoft Graph:

Convert to PDF (format query parameter):
Convert endpoint doc: https://learn.microsoft.com/graph/api/driveitem-get-content-format?view=graph-rest-1.0
(Example: GET /drives/{drive-id}/items/{item-id}/content?format=pdf)

Download original file (if you need to pre-check or compare):
https://learn.microsoft.com/graph/api/driveitem-get-content?view=graph-rest-1.0

Java SDK resources (updated):
SDK overview: https://learn.microsoft.com/graph/sdks/sdks-overview#java
Delegated auth tutorial: https://learn.microsoft.com/graph/tutorials/java
App-only auth tutorial: https://learn.microsoft.com/graph/tutorials/java-app-only
SDK repo: https://github.com/microsoftgraph/msgraph-sdk-java

Before investing more time tuning thee cloud conversion, try one desktop Word export (with the same fonts installed) and compare the PDFs. That concrete difference, if any, will tell you whether a controlled desktop path is worth operational overhead.

Best regards,
Raymond
Naresh Namani 0 Reputation points

2025-09-30T09:28:47.0266667+00:00

Hi Raymond,

you were saying that MS Graph ,Word Online(power automate) , Office 365 apis uses same rendering engine ? if not ,can you please provide documentation of Office 365 apis ?

i think word desktop application is not viable approach because our java application runs on linux environment and MS word desktop runs only on Windows , platform dependent .

we have checked exporting from MS word Desktop , pdf and .docx files looks same. Online MS graph , Word Business (online - Power Automate ) is generated pdf inaccurately.

can you please provide clarity on office 365 api , MS graph , word online for business (power automate)

Thank you.
Anonymous

2025-10-01T08:20:17.66+00:00
Hi Naresh Namani ,

Thanks for your detailed follow-up, let me try to clarify:

Microsoft Graph, Word Online (Power Automate), and the Office 365 conversion APIs all rely on a cloud-based rendering engine, which is different from the one used by the Word desktop client. The desktop application uses the full native Word rendering engine, so exported PDFs will match exactly what you see in the Word editor. The cloud rendering engine is optimized for scalability and online scenarios, but layout fidelity can differ in areas such as pagination, line breaks, and placement of complex objects.

Because your Java application runs on Linux, using Word desktop directly is not feasible. That said, here are some references you can check for each API/service:

Microsoft Graph DriveItem conversion endpoint:

https://learn.microsoft.com/en-us/graph/api/driveitem-get-content?view=graph-rest-1.0&tabs=http#example-convert-to-pdf

Graph SDK for Java:

https://learn.microsoft.com/en-us/graph/sdks/sdks-overview

Word Online (Business) / Power Automate Word connector:

https://learn.microsoft.com/en-us/connectors/wordonlinebusiness/

To answer your specific point: yes, the online services (Graph conversion, Word Online connector, Power Automate actions) share the same cloud rendering engine, and this is why you see similar output issues when comparing them. None of these APIs guarantee identical pagination to Word desktop.

If exact fidelity is a strict requirement, the only fully reliable option is automating Word desktop (for example in a Windows VM or server environment). Otherwise, you can mitigate differences by standardizing fonts, page sizes, margins, and avoiding complex floating elements in your documents, but results may still vary.

Hope this clears it up.

Best regards,

Raymond
Anonymous

2025-10-06T10:46:58.9866667+00:00

Hi ,

Just checking in to see if the above information was helpful. If you have any further updates on this issue, please feel free to post back.

Share via

Graph API - Docx to Pdf Conversion .

1 answer

Your answer