[GH-ISSUE #10852] add toggle to disable rag or store embeding into vector db for customized pipeline when attaching documents #86793

Closed
opened 2026-05-15 11:07:10 -05:00 by GiteaMirror · 0 comments
Owner

Originally created by @wangjiyang on GitHub (Feb 27, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/10852

Feature Request


Is your feature request related to a problem? Please describe.
I’m using Open WebUI to upload attachments of various sizes (sometimes very large). Currently, when an attachment is uploaded, Open WebUI automatically converts the file into text and attempts to store it in a vector database. This can be problematic for very large files, as storing massive amounts of data in a vector database might be impractical or even impossible. I also prefer handling the file processing separately in my pipeline, so I don’t always need the built-in RAG (Retrieval-Augmented Generation) or vector storage features.

For example, if I upload a 2GB text file, Open WebUI tries to parse and store it in the vector database by default. This is both time-consuming and storage-intensive. Moreover, it doesn’t align with the way my current pipeline processes attachments.


Describe the solution you’d like
I would like an option (e.g., a toggle in the settings or a command-line flag) that completely disables or bypasses the RAG-based text extraction and vector database storage steps. Ideally, this setting could be turned on or off depending on the user’s workflow:

  1. On – The current behavior remains as is: any uploaded attachment is automatically processed and saved in the vector database.

  2. Off – Attachments are not converted into text and are not stored in the vector database. Instead, the user can handle file processing independently in their own pipeline.


Describe alternatives you’ve considered

Manual Removal of Vector Storage Code: Commenting out or removing the relevant code segments in the Open WebUI source. However, this is not ideal since future updates or merges might overwrite these modifications.

External Scripts: Using external scripts to intercept uploaded files before they reach Open WebUI. This allows me to store or process them elsewhere, but it doesn’t fully prevent Open WebUI from attempting to vectorize large files.


Additional context
An example scenario:

  1. I upload a large PDF or text file (exceeding 1GB).

  2. Open WebUI automatically begins parsing the file text and storing it into the vector database. This can lead to resource exhaustion, reduced performance, or even a crash if the file is too large.

  3. My separate pipeline is already optimized to process large files outside of Open WebUI, so I don’t need or want the built-in text conversion step.

A simple toggle or setting to disable RAG/vector storage would streamline workflows that process large attachments differently. This feature would make Open WebUI more flexible and better suited for a wider range of real-world use cases.

Originally created by @wangjiyang on GitHub (Feb 27, 2025). Original GitHub issue: https://github.com/open-webui/open-webui/issues/10852 Feature Request --- Is your feature request related to a problem? Please describe. I’m using Open WebUI to upload attachments of various sizes (sometimes very large). Currently, when an attachment is uploaded, Open WebUI automatically converts the file into text and attempts to store it in a vector database. This can be problematic for very large files, as storing massive amounts of data in a vector database might be impractical or even impossible. I also prefer handling the file processing separately in my pipeline, so I don’t always need the built-in RAG (Retrieval-Augmented Generation) or vector storage features. For example, if I upload a 2GB text file, Open WebUI tries to parse and store it in the vector database by default. This is both time-consuming and storage-intensive. Moreover, it doesn’t align with the way my current pipeline processes attachments. --- Describe the solution you’d like I would like an option (e.g., a toggle in the settings or a command-line flag) that completely disables or bypasses the RAG-based text extraction and vector database storage steps. Ideally, this setting could be turned on or off depending on the user’s workflow: 1. On – The current behavior remains as is: any uploaded attachment is automatically processed and saved in the vector database. 2. Off – Attachments are not converted into text and are not stored in the vector database. Instead, the user can handle file processing independently in their own pipeline. --- Describe alternatives you’ve considered Manual Removal of Vector Storage Code: Commenting out or removing the relevant code segments in the Open WebUI source. However, this is not ideal since future updates or merges might overwrite these modifications. External Scripts: Using external scripts to intercept uploaded files before they reach Open WebUI. This allows me to store or process them elsewhere, but it doesn’t fully prevent Open WebUI from attempting to vectorize large files. --- Additional context An example scenario: 1. I upload a large PDF or text file (exceeding 1GB). 2. Open WebUI automatically begins parsing the file text and storing it into the vector database. This can lead to resource exhaustion, reduced performance, or even a crash if the file is too large. 3. My separate pipeline is already optimized to process large files outside of Open WebUI, so I don’t need or want the built-in text conversion step. A simple toggle or setting to disable RAG/vector storage would streamline workflows that process large attachments differently. This feature would make Open WebUI more flexible and better suited for a wider range of real-world use cases.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/open-webui#86793