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 #5250] [BUG]: If agent web search is not setup at ALLM level, the system goes into infinite loop. #4985

New Issue
Closed
opened 2026-06-05 14:51:18 -04:00 by yindo · 2 comments
yindo commented 2026-06-05 14:51:18 -04:00
Owner
Copy Link
Copy Source

Originally created by @bulrush15 on GitHub (Mar 22, 2026).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5250

How are you running AnythingLLM?

AnythingLLM desktop app

What happened?

  • My system: HP Omen GT13-1xxx desktop PC with Windows 11, 16GBM RAM, UEFI
  • CPU: AMD Ryzen 5 5600G with Radeon Graphics, 6 cores
  • Video card: Nvidia GeForce RTX 3060 with 12GBM VRAM

The agent should fail gracefully. It should check to see 1) if the web search (or other agent tool) is on and then 2) Check to see if a web search provider was given in the system setup, or workspace setup.

In AnythingLLM Desktop v1.11.2 I didn't have my agent web search on or set up.

Are there known steps to reproduce?

  1. Install a fresh version of AnythingLLM without setting up the web search tool for agents.
  2. I typed "@agent find the top 8 articles about AI and search the web. List the articles one per bullet and have a clickable link in each bullet for the source".

The agent seemed to start but never stopped eventually using over 80% of my memory. When I stopped ALLM via Task Manager the memory was never freed. I had to reboot the PC.

Originally created by @bulrush15 on GitHub (Mar 22, 2026). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5250 ### How are you running AnythingLLM? AnythingLLM desktop app ### What happened? * My system: HP Omen GT13-1xxx desktop PC with Windows 11, 16GBM RAM, UEFI * CPU: AMD Ryzen 5 5600G with Radeon Graphics, 6 cores * Video card: Nvidia GeForce RTX 3060 with 12GBM VRAM The agent should fail gracefully. It should check to see 1) if the web search (or other agent tool) is on and then 2) Check to see if a web search provider was given in the system setup, or workspace setup. In AnythingLLM Desktop v1.11.2 I didn't have my agent web search on or set up. ### Are there known steps to reproduce? 1. Install a fresh version of AnythingLLM without setting up the web search tool for agents. 2. I typed "@agent find the top 8 articles about AI and search the web. List the articles one per bullet and have a clickable link in each bullet for the source". The agent seemed to start but never stopped eventually using over 80% of my memory. When I stopped ALLM via Task Manager the memory was never freed. I had to reboot the PC.
yindo added the possible bugneeds info / can't replicate labels 2026-06-05 14:51:18 -04:00
yindo closed this issue 2026-06-05 14:51:19 -04:00
yindo commented 2026-06-05 14:51:19 -04:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Mar 23, 2026):

The agent seemed to start but never stopped eventually using over 80% of my memory. When I stopped ALLM via Task Manager the memory was never freed. I had to reboot the PC.

I dont know what llm provider you are using but killing AnythingLLM would indeed kill the child llm.exe that was associated with the parent process.

We have a max tool call stack to prevent exactly this with small models that are not "smart" enough to know when to stop calling tools - which is common for some providers who do not support native tool calling.

Are you using the default llm provider and what specific model are you on?

<!-- gh-comment-id:4111990505 --> @timothycarambat commented on GitHub (Mar 23, 2026): > The agent seemed to start but never stopped eventually using over 80% of my memory. When I stopped ALLM via Task Manager the memory was never freed. I had to reboot the PC. I dont know what llm provider you are using but killing AnythingLLM would indeed kill the child `llm.exe` that was associated with the parent process. We have a max tool call stack to prevent exactly this with small models that are not "smart" enough to know when to stop calling tools - which is common for some providers who do not support native tool calling. Are you using the default llm provider and what specific model are you on?
yindo commented 2026-06-05 14:51:19 -04:00
Author
Owner
Copy Link
Copy Source

@bulrush15 commented on GitHub (Mar 25, 2026):

That workspace is using llama3.2:3b. I didn't realize this could be a model issue. There was no timeout, the ALLM app just ran and ran with no message at all.

<!-- gh-comment-id:4125758878 --> @bulrush15 commented on GitHub (Mar 25, 2026): That workspace is using llama3.2:3b. I didn't realize this could be a model issue. There was no timeout, the ALLM app just ran and ran with no message at all.
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 needs info / can't replicate 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#4985
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?