mirror of
https://github.com/open-webui/open-webui.git
synced 2026-07-11 18:43:44 -05:00
Web browsing using # isn't scraping content #2378
No Assignees
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: github-starred/open-webui#2378
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @rfernandez760 on GitHub (Oct 14, 2024).
Bug Report
Important: Before submitting a bug report, please check whether a similar issue or feature request has already been posted in the Issues or Discussions section. It's likely we're already tracking it. In case of uncertainty, initiate a discussion post first. This helps us all to efficiently focus on improving the project.
Let's collaborate respectfully. If you bring negativity, please understand our capacity to engage may be limited. If you're open to learning and communicating constructively, we're more than happy to assist you. Remember, Open WebUI is a volunteer-driven project maintained by a single maintainer, supported by our amazing contributors who also manage full-time jobs. We respect your time; please respect ours. If you have an issue, We highly encourage you to submit a pull request or to fork the project. We actively work to prevent contributor burnout to preserve the quality and continuity of Open WebUI.
Installation Method
OWUI Installed via Docker
Environment
Open WebUI Version: v0.3.32
Ollama (if applicable): 0.3.13
Operating System: Win11, Running WSL2 with Ubuntu
Browser (if applicable): Chrome and Edge
Confirmation:
Expected Behavior:
Scenario 1. When adding in # at the beginning or in the message, OWUI will generate a file for the LLM review much like attaching a file that has been uploaded for RAG.
Scenario 2. When adding # in the message, OWUI will scrape the content of the web page for RAG. Regardless of the type of LLM model selected.
Actual Behavior:
Scenario 1. The model is not able to retrieve any data. The citation just shows the URL link that is provided.
Scenario 2. Closed source models like Claude or GPT4.0 can't read the web page, but open source models can. Please note the open source models tested were llama3.1 that is installed locally as well as open source models connected to Groq via API.
Description
Bug Summary:
Web browsing (ability to read a web page) is unavailable, and inconsistent behavior based on open source vs closed source models. This I don't believe is the expected functionality.
Reproduction Details
Steps to Reproduce:
Scenario 1: Add a # first in the chat so OWUI can generate the document (content) for review. Then ask any questions like summarize. Use any model open source or closed source.
Scenario 2: Type in the message field "Please summarize the following #". Ask using an open source, and then create a new message following the same steps using a closed source model. The open is able to retrieve and read, while the closed source doesn't have any context.
Example URL: #https://www.msn.com/en-us/news/politics/cruz-calls-out-mcconnell-aligned-group-for-not-spending-a-penny-in-texas/ar-AA1sgaor?ocid=msedgntp&pc=U531&cvid=8e102387764f4ce3925621b1d3a529c0&ei=10
Logs and Screenshots
Browser Console Logs:
[Include relevant browser console logs, if applicable]
Docker Container Logs:
[Include relevant Docker container logs, if applicable]
Docker Logs.txt
Screenshots/Screen Recordings (if applicable):
Additional Information
[Include any additional details that may help in understanding and reproducing the issue. This could include specific configurations, error messages, or anything else relevant to the bug.]
Note
If the bug report is incomplete or does not follow the provided instructions, it may not be addressed. Please ensure that you have followed the steps outlined in the README.md and troubleshooting.md documents, and provide all necessary information for us to reproduce and address the issue. Thank you!