Mintplex-Labs/anything-llm
1
0
Fork 0
mirror of https://github.com/Mintplex-Labs/anything-llm.git synced 2026-07-19 14:13:35 -04:00
Code Issues 328 Packages Projects Releases 10 Wiki Activity

[GH-ISSUE #2208] [BUG]: Agent with webscraping is hallucinating #1443

New Issue
Closed
opened 2026-02-22 18:24:49 -05:00 by yindo · 3 comments
yindo commented 2026-02-22 18:24:49 -05:00
Owner
Copy Link
Copy Source

Originally created by @jeremiasbaur on GitHub (Sep 2, 2024).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/2208

How are you running AnythingLLM?

AnythingLLM desktop app

What happened?

I discovered a bug while using the agent with the web-scraping command.

I was trying to scrape multiple lecture pages to generate a structured format so I can paste the content into an excel file. I instructed my agent in what information I am interested in and then started it with the first webpage. For the consecutive webpages, I was using this command:

@agent
Do the same for:
@web-scraping www.website.com

Sometimes, the agent correctly accessed the webpage and gave me the correct information. However, if I instantly sent the next command with the next website, the agent didn't scrape the website and just hallucinated some lecture information. If I only send the command after e.g. 30s, everything works well. I know that I don't need to use @agent every time but it shouldn't change anything, right?

Why does this happen and how can you / I fix it?

Are there known steps to reproduce?

No response

Originally created by @jeremiasbaur on GitHub (Sep 2, 2024). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/2208 ### How are you running AnythingLLM? AnythingLLM desktop app ### What happened? I discovered a bug while using the agent with the web-scraping command. I was trying to scrape multiple lecture pages to generate a structured format so I can paste the content into an excel file. I instructed my agent in what information I am interested in and then started it with the first webpage. For the consecutive webpages, I was using this command: @agent Do the same for: @web-scraping www.website.com Sometimes, the agent correctly accessed the webpage and gave me the correct information. However, if I instantly sent the next command with the next website, the agent didn't scrape the website and just hallucinated some lecture information. If I only send the command after e.g. 30s, everything works well. I know that I don't need to use @agent every time but it shouldn't change anything, right? Why does this happen and how can you / I fix it? ### Are there known steps to reproduce? _No response_
yindo added the possible bug label 2026-02-22 18:24:49 -05:00
yindo closed this issue 2026-02-22 18:24:49 -05:00
yindo commented 2026-02-22 18:24:49 -05:00
Author
Owner
Copy Link
Copy Source

@pleabargain commented on GitHub (Sep 2, 2024):

Which v. of the desktop are you using? OS?

@pleabargain commented on GitHub (Sep 2, 2024): Which v. of the desktop are you using? OS?
yindo commented 2026-02-22 18:24:49 -05:00
Author
Owner
Copy Link
Copy Source

@jeremiasbaur commented on GitHub (Sep 2, 2024):

I am using the version 1.6.4

@jeremiasbaur commented on GitHub (Sep 2, 2024): I am using the version 1.6.4
yindo commented 2026-02-22 18:24:49 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Sep 3, 2024):

When the @agent session closes it does not keep the previous session data (like an entire website) in memory. If you want to persist the data of a whole website that was scraped use the actual website scraper and embed it is a document. Ephemeral data an agent accumulates during a session is currently not persisted between sessions since it may not be useful after the session is completed.

If you dont send a chat for 30s, the session can close and the UI will report this as agent session ended in which sending a chat afterwards is like using regular rag.

@timothycarambat commented on GitHub (Sep 3, 2024): When the `@agent` session closes it does not keep the previous session data (like an entire website) in memory. If you want to persist the data of a whole website that was scraped use the actual website scraper and embed it is a document. Ephemeral data an agent accumulates during a session is currently not persisted between sessions since it may not be useful after the session is completed. If you dont send a chat for 30s, the session can close and the UI will report this as `agent session ended` in which sending a chat afterwards is like using regular rag.
yindo changed title from [BUG]: Agent with webscraping is hallucinating to [GH-ISSUE #2208] [BUG]: Agent with webscraping is hallucinating 2026-06-05 14:40:48 -04:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
Desktop
Docker
Integration Request
Integration Request
OS: Linux
OS: Mobile
OS: Windows
UI/UX
blocked
bug
bug
core-team-only
documentation
duplicate
embed-widget
enhancement
feature request
github_actions
good first issue
investigating
needs info / can't replicate
possible bug
pull-request
Mirrored from GitHub Pull Request
 question
stage: specifications
wontfix
No Label possible bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Mintplex-Labs/anything-llm#1443
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?