[GH-ISSUE #18144] feat: document upload available raw for other usages #18510

Closed
opened 2026-04-20 00:44:37 -05:00 by GiteaMirror · 4 comments
Owner

Originally created by @MarceloMassarente on GitHub (Oct 8, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/18144

Check Existing Issues

  • I have searched the existing issues and discussions.

Problem Description

It would be great if when I upload a pdf or csv file it is available raw and not only processed by RAG.
for example, to send a raw pdf file to Claude API or Gemini for their vision capabilities to apply, or to have csv file available to be sent to Jupyter, or to apply a custom RAG/Doc processing tool on the file/collection

Desired Solution you'd like

add an "raw upload" button additionally to the current file upload that bypasses RAG and keep full doc available

Alternatives Considered

No response

Additional Context

No response

Originally created by @MarceloMassarente on GitHub (Oct 8, 2025). Original GitHub issue: https://github.com/open-webui/open-webui/issues/18144 ### Check Existing Issues - [x] I have searched the existing issues and discussions. ### Problem Description It would be great if when I upload a pdf or csv file it is available raw and not only processed by RAG. for example, to send a raw pdf file to Claude API or Gemini for their vision capabilities to apply, or to have csv file available to be sent to Jupyter, or to apply a custom RAG/Doc processing tool on the file/collection ### Desired Solution you'd like add an "raw upload" button additionally to the current file upload that bypasses RAG and keep full doc available ### Alternatives Considered _No response_ ### Additional Context _No response_
Author
Owner

@rohithshenoyg commented on GitHub (Oct 8, 2025):

I may not have fully understood your request, but after the file is uploaded, it can be downloaded in its original form by clicking any citation from the file and the title of the citation dialog box where the file name is present "something.csv" can be clicked to download.

<!-- gh-comment-id:3381984442 --> @rohithshenoyg commented on GitHub (Oct 8, 2025): I may not have fully understood your request, but after the file is uploaded, it can be downloaded in its original form by clicking any citation from the file and the title of the citation dialog box where the file name is present "something.csv" can be clicked to download.
Author
Owner

@silentoplayz commented on GitHub (Oct 8, 2025):

Is this what you're wanting? Try toggling Bypass Embedding and Retrieval on in the admin Documents settings and see if that produces the result you want.

Image
<!-- gh-comment-id:3382092838 --> @silentoplayz commented on GitHub (Oct 8, 2025): Is this what you're wanting? Try toggling `Bypass Embedding and Retrieval` on in the admin `Documents` settings and see if that produces the result you want. <img width="2165" height="403" alt="Image" src="https://github.com/user-attachments/assets/8d9efb00-ea2a-4198-818a-5c866d9e48de" />
Author
Owner

@MarceloMassarente commented on GitHub (Oct 8, 2025):

I am sorry I didn't explain well.
I don't want to download it, but use it in Openwebui in its raw format.
Imagine I have a PDF and want to send it as PDF to Claude for processing, not the pre processed text generated by the back end.
When I upload it to OWUI today, it is always processed and not send in full with the images, etc
I would like a way to send it the same way I send a jpg when added to the chat

<!-- gh-comment-id:3382099508 --> @MarceloMassarente commented on GitHub (Oct 8, 2025): I am sorry I didn't explain well. I don't want to download it, but use it in Openwebui in its raw format. Imagine I have a PDF and want to send it as PDF to Claude for processing, not the pre processed text generated by the back end. When I upload it to OWUI today, it is always processed and not send in full with the images, etc I would like a way to send it the same way I send a jpg when added to the chat
Author
Owner

@rgaricano commented on GitHub (Oct 8, 2025):

You can use a filter function, setting self.file_handler = True in a filter's __init__ method, it tells Open WebUI to defer file-related operations to your filter instead of using the default content extraction pipeline.

e.g. of filter function for do that:

Skip_Content_Extraction_Filter.py

<!-- gh-comment-id:3382959537 --> @rgaricano commented on GitHub (Oct 8, 2025): You can use a filter function, setting `self.file_handler = True` in a filter's `__init__` method, it tells Open WebUI to defer file-related operations to your filter instead of using the default content extraction pipeline. e.g. of filter function for do that: [Skip_Content_Extraction_Filter.py](https://github.com/user-attachments/files/22784686/Skip_Content_Extraction_Filter.py)
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/open-webui#18510