Mintplex-Labs/anything-llm
1
0
Fork 0
mirror of https://github.com/Mintplex-Labs/anything-llm.git synced 2026-07-19 22:23:50 -04:00
Code Issues 328 Packages Projects Releases 10 Wiki Activity

[GH-ISSUE #3903] Loss of ollama models when AnythingLLM auto discovers Ollama provider #2488

New Issue
Closed
opened 2026-02-22 18:29:54 -05:00 by yindo · 3 comments
yindo commented 2026-02-22 18:29:54 -05:00
Owner
Copy Link
Copy Source

Originally created by @dermotholmes on GitHub (May 28, 2025).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/3903

How are you running AnythingLLM?

AnythingLLM desktop app

What happened?

Went to try out AnythingLLM Desktop (OSX 15) yesterday. Got as far as opening the app and setting Ollama as a provider (which it successfully discovered) before being distracted by other tasks.

Later (for other purposes) ran ollama list to find that no models were listed 🫨 - I had used multiple models in ollama the day before. I checked if the model files (blobs and manifests folders) were actually still present in /Users/user/.ollama - yes they are.

Pulled a new model from Ollama with ollama pull model_x and found that this model had actually been saved into: '/Users/user/Library/Application Support/anythingllm-desktop/storage/models/ollama'

It appears the installation of Anything LLM has changed the location that Ollama is saving and looking for models in? Is this expected?

EDIT I noticed that this is only a problem when Anything LLM is running. When Anything LLM is closed, Ollama reverts back to using it's original directory!

Notes

  • Run on OSX 15.5
  • As I'm using the Ollama desktop app I'm not sure where/how to check any path variables that may have been changed.

Are there known steps to reproduce?

  1. Install Ollama (Mac app version)
  2. Pull any model (model A) to Ollama
  3. Run ollama list to verify model A is downloaded
  4. Install Anything LLM
  5. Run Anything LLM
  6. Setup Ollama as a provider in Anything LLM
  7. Run ollama list > the downloaded model A should be missing (Anything LLM MUST be running while this is done)
  8. Pull another model (model X) via ollama
  9. verify model X downloaded with ollama list
  10. verify model X downloaded to location '/Users/user/Library/Application Support/anythingllm-desktop/storage/models/ollama/...'
Originally created by @dermotholmes on GitHub (May 28, 2025). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/3903 ### How are you running AnythingLLM? AnythingLLM desktop app ### What happened? Went to try out AnythingLLM Desktop (OSX 15) yesterday. Got as far as opening the app and setting Ollama as a provider (which it successfully discovered) before being distracted by other tasks. Later (for other purposes) ran `ollama list` to find that no models were listed 🫨 - I had used multiple models in ollama the day before. I checked if the model files (blobs and manifests folders) were actually still present in /Users/user/.ollama - yes they are. Pulled a new model from Ollama with `ollama pull model_x` and found that this model had actually been saved into: '/Users/user/Library/Application Support/anythingllm-desktop/storage/models/ollama' It appears the installation of Anything LLM has changed the location that Ollama is saving and looking for models in? Is this expected? *EDIT* I noticed that this is only a problem when Anything LLM is running. When Anything LLM is closed, Ollama reverts back to using it's original directory! --- Notes - Run on OSX 15.5 - As I'm using the Ollama desktop app I'm not sure where/how to check any path variables that may have been changed. ### Are there known steps to reproduce? 1. Install Ollama (Mac app version) 2. Pull any model (model A) to Ollama 3. Run `ollama list` to verify model A is downloaded 4. Install Anything LLM 5. Run Anything LLM 6. Setup Ollama as a provider in Anything LLM 7. Run `ollama list` > the downloaded model A should be missing (Anything LLM MUST be running while this is done) 8. Pull another model (model X) via ollama 9. verify model X downloaded with `ollama list` 10. verify model X downloaded to location '/Users/user/Library/Application Support/anythingllm-desktop/storage/models/ollama/...'
yindo added the possible bugDesktop labels 2026-02-22 18:29:54 -05:00
yindo closed this issue 2026-02-22 18:29:54 -05:00
yindo commented 2026-02-22 18:29:55 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (May 28, 2025):

Does this happen if Ollama (on host machine) is running then you open AnythingLLM and use the default embedder?

@timothycarambat commented on GitHub (May 28, 2025): Does this happen if Ollama (on host machine) is running _then_ you open AnythingLLM and use the default embedder?
yindo commented 2026-02-22 18:29:55 -05:00
Author
Owner
Copy Link
Copy Source

@dermotholmes commented on GitHub (May 28, 2025):

Just tested:

  • When Ollama (launched via mac app) is running -> Then no, the issue is not present and ollama list returns my models as expected

I also tested the inverse (Ollama not running -> Open Anything LLM) and that is not a problem either.

I can't seem to be able to replicate the issue anymore :/

The only other comments I have that may provide hints are:

  1. I've restarted my machine between filing the issue and now
  2. When I encountered the issue, I hadn't made it through the setup screens in Anything LLM i.e. the app state was on one of those setup screens for about a day while I did other things - mentioning that incase there's any config logic that takes places during those steps.

Happy to test any other scenarios if you need! Otherwise feel free to close and I'll re-raise if it crops up again.

@dermotholmes commented on GitHub (May 28, 2025): Just tested: - When Ollama (launched via mac app) is running -> Then no, the issue is not present and `ollama list` returns my models as expected I also tested the inverse (Ollama not running -> Open Anything LLM) and that is not a problem either. I can't seem to be able to replicate the issue anymore :/ The only other comments I have that may provide hints are: 1. I've restarted my machine between filing the issue and now 2. When I encountered the issue, I hadn't made it through the setup screens in Anything LLM i.e. the app state was on one of those setup screens for about a day while I did other things - mentioning that incase there's any config logic that takes places during those steps. Happy to test any other scenarios if you need! Otherwise feel free to close and I'll re-raise if it crops up again.
yindo commented 2026-02-22 18:29:55 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Jun 3, 2025):

Thank you for at least documenting that - lets just close and see if it becomes a recurring issue and address from there since it isnt easily or consistently replicable for now

@timothycarambat commented on GitHub (Jun 3, 2025): Thank you for at least documenting that - lets just close and see if it becomes a recurring issue and address from there since it isnt easily or consistently replicable for now
yindo changed title from Loss of ollama models when AnythingLLM auto discovers Ollama provider to [GH-ISSUE #3903] Loss of ollama models when AnythingLLM auto discovers Ollama provider 2026-06-05 14:46:50 -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 Desktop 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#2488
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?