[GH-ISSUE #16469] issue: With "Using Entire Document", document is not "Entire" #33448

Closed
opened 2026-04-25 07:22:13 -05:00 by GiteaMirror · 3 comments
Owner

Originally created by @chdid on GitHub (Aug 11, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/16469

Check Existing Issues

  • I have searched the existing issues and discussions.
  • I am using the latest version of Open WebUI.

Installation Method

Docker

Open WebUI Version

0.6.21

Ollama Version (if applicable)

No response

Operating System

ubuntu

Browser (if applicable)

No response

Confirmation

  • I have read and followed all instructions in README.md.
  • I am using the latest version of both Open WebUI and Ollama.
  • I have included the browser console logs.
  • I have included the Docker container logs.
  • I have provided every relevant configuration, setting, and environment variable used in my setup.
  • I have clearly listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
  • I have documented step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation. My steps:
  • Start with the initial platform/version/OS and dependencies used,
  • Specify exact install/launch/configure commands,
  • List URLs visited, user input (incl. example values/emails/passwords if needed),
  • Describe all options and toggles enabled or changed,
  • Include any files or environmental changes,
  • Identify the expected and actual result at each stage,
  • Ensure any reasonably skilled user can follow and hit the same issue.

Expected Behavior

Have full document consideration when I check "Using Entire Document"

Actual Behavior

The document is partially processed.

Steps to Reproduce

Take a file with 800 entries.
Uncheck the "Paste large text as a file" box in the user settings.

Nominal OK case:

  • copy the table with 800 entries
  • paste the table into the prompt
  • ask for the number of entries
  • you should get a result like "the table has 800 entries"

Unknown case
Check the "Paste large text as a file" box in the user settings.

  • Generate a new conversation
  • copy the table (it is transformed into an attached file with "Using Entire Document" checked).
  • ask for the number of entries
  • you should get a result like "the table has 290 entries"

The result is the same if we add the file manually and then check "Using Entire Document".

Logs & Screenshots

No problem in logs.

Documents in settings
Image

Additional Information

No response

Originally created by @chdid on GitHub (Aug 11, 2025). Original GitHub issue: https://github.com/open-webui/open-webui/issues/16469 ### Check Existing Issues - [x] I have searched the existing issues and discussions. - [x] I am using the latest version of Open WebUI. ### Installation Method Docker ### Open WebUI Version 0.6.21 ### Ollama Version (if applicable) _No response_ ### Operating System ubuntu ### Browser (if applicable) _No response_ ### Confirmation - [x] I have read and followed all instructions in `README.md`. - [x] I am using the latest version of **both** Open WebUI and Ollama. - [x] I have included the browser console logs. - [x] I have included the Docker container logs. - [x] I have **provided every relevant configuration, setting, and environment variable used in my setup.** - [x] I have clearly **listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup** (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc). - [x] I have documented **step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation**. My steps: - Start with the initial platform/version/OS and dependencies used, - Specify exact install/launch/configure commands, - List URLs visited, user input (incl. example values/emails/passwords if needed), - Describe all options and toggles enabled or changed, - Include any files or environmental changes, - Identify the expected and actual result at each stage, - Ensure any reasonably skilled user can follow and hit the same issue. ### Expected Behavior Have full document consideration when I check "Using Entire Document" ### Actual Behavior The document is partially processed. ### Steps to Reproduce Take a file with 800 entries. Uncheck the "Paste large text as a file" box in the user settings. Nominal OK case: - copy the table with 800 entries - paste the table into the prompt - ask for the number of entries - you should get a result like "the table has 800 entries" Unknown case Check the "Paste large text as a file" box in the user settings. - Generate a new conversation - copy the table (it is transformed into an attached file with "Using Entire Document" checked). - ask for the number of entries - you should get a result like "the table has 290 entries" The result is the same if we add the file manually and then check "Using Entire Document". ### Logs & Screenshots No problem in logs. Documents in settings <img width="723" height="445" alt="Image" src="https://github.com/user-attachments/assets/eeb16525-6d68-4dda-9ad7-785092334a62" /> ### Additional Information _No response_
GiteaMirror added the bug label 2026-04-25 07:22:13 -05:00
Author
Owner

@frenzybiscuit commented on GitHub (Aug 11, 2025):

You realize LLM's can't count hundreds of entries, right?

Does this work for you outside this platform?

<!-- gh-comment-id:3173617567 --> @frenzybiscuit commented on GitHub (Aug 11, 2025): You realize LLM's can't count hundreds of entries, right? Does this work for you outside this platform?
Author
Owner

@chdid commented on GitHub (Aug 11, 2025):

That's what I initially thought.

I realize I need to be more specific with the example.

I'm using the qwen3 model.

The table used includes a list of 797 PCs with the name and type of PC.

The prompt is: "Count the number of PCs by type
The number of PCs is 797. Check the transmitted data."

If the table is pasted into the prompt, the count is OK.

If the table is attached, it only finds 290 or a similar number.

The LLM is indeed having difficulty processing so much data. What surprises me here is the difference in response between the version with the file and the copy-paste version into the prompt.

<!-- gh-comment-id:3173877637 --> @chdid commented on GitHub (Aug 11, 2025): That's what I initially thought. I realize I need to be more specific with the example. I'm using the qwen3 model. The table used includes a list of 797 PCs with the name and type of PC. The prompt is: "Count the number of PCs by type The number of PCs is 797. Check the transmitted data." If the table is pasted into the prompt, the count is OK. If the table is attached, it only finds 290 or a similar number. The LLM is indeed having difficulty processing so much data. What surprises me here is the difference in response between the version with the file and the copy-paste version into the prompt.
Author
Owner

@tjbck commented on GitHub (Aug 11, 2025):

#16474

<!-- gh-comment-id:3174005391 --> @tjbck commented on GitHub (Aug 11, 2025): #16474
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/open-webui#33448