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 #5149] AnythingLLM is not local. #4934

New Issue
Closed
opened 2026-06-05 14:50:58 -04:00 by yindo · 10 comments
yindo commented 2026-06-05 14:50:58 -04:00
Owner
Copy Link
Copy Source

Originally created by @Arifursdev on GitHub (Mar 5, 2026).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5149

Don't know how offline this is but I've installed this to try it out. it suggested to download a specific qwen3 vision model, downloaded it and started chatting thinking it was all local. it was responding within milliseconds, and was checking task manager that its actually using internet whenever i'm sending a message in chat.

so I just quickly unplugged ethernet cable, and voila. in the chat its just loading.

so my question is how is this local and if its local why did it download the model first and starting sending my data to a cloud server from china?

Originally created by @Arifursdev on GitHub (Mar 5, 2026). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5149 Don't know how offline this is but I've installed this to try it out. it suggested to download a specific qwen3 vision model, downloaded it and started chatting thinking it was all local. it was responding within milliseconds, and was checking task manager that its actually using internet whenever i'm sending a message in chat. so I just quickly unplugged ethernet cable, and voila. in the chat its just loading. so my question is how is this local and if its local why did it download the model first and starting sending my data to a cloud server from china?
yindo closed this issue 2026-06-05 14:50:59 -04:00
yindo commented 2026-06-05 14:50:59 -04:00
Author
Owner
Copy Link
Copy Source

@DE4DLY commented on GitHub (Mar 5, 2026):

There might be some telemery... but i doubt it isn't anonymous and stated in the project. Pull the chord or turn off wifi... and it still works....

I'm inspecting my network trafic and i have no trafic coming and going from ALLM on my Unifi network console.

Image

@timothycarambat would be the best to answer this. If you're concerned, just learn to self-host an Ollama server or LM-Studio for total control over the LLM you use and the data coming in and out of them.

edit"

here... in the readme...

Image
<!-- gh-comment-id:4005851663 --> @DE4DLY commented on GitHub (Mar 5, 2026): There might be some telemery... but i doubt it isn't anonymous and stated in the project. Pull the chord or turn off wifi... and it still works.... I'm inspecting my network trafic and i have no trafic coming and going from ALLM on my Unifi network console. <img width="984" height="360" alt="Image" src="https://github.com/user-attachments/assets/76099305-8b3c-42b8-9215-548dd33a1b21" /> @timothycarambat would be the best to answer this. If you're concerned, just learn to self-host an Ollama server or LM-Studio for total control over the LLM you use and the data coming in and out of them. **edit**" here... in the readme... <img width="1033" height="907" alt="Image" src="https://github.com/user-attachments/assets/877dc733-c6c6-4118-bb68-8e8dbeef0e15" />
yindo commented 2026-06-05 14:50:59 -04:00
Author
Owner
Copy Link
Copy Source

@Arifursdev commented on GitHub (Mar 5, 2026):

not so "fully local and offline"..
Image

It would have been fine but what's suspicious is, it downloaded some qwen3 vision model around 5.5~ GB and started to chat with a cloud AI of qwen.. stopped giving response once I unplugged the ethernet.

@DE4DLY if it was telemetry, It wouldn't stop responding to the chat, right? remember model is downloaded and its "fully local, offline"

<!-- gh-comment-id:4005919606 --> @Arifursdev commented on GitHub (Mar 5, 2026): not so "fully local and offline".. <img width="1146" height="294" alt="Image" src="https://github.com/user-attachments/assets/ac0f61e5-f821-4a5a-8e5b-24f6a69731cd" /> It would have been fine but what's suspicious is, it downloaded some qwen3 vision model around 5.5~ GB and started to chat with a cloud AI of qwen.. stopped giving response once I unplugged the ethernet. @DE4DLY if it was telemetry, It wouldn't stop responding to the chat, right? remember model is downloaded and its "fully local, offline"
yindo commented 2026-06-05 14:50:59 -04:00
Author
Owner
Copy Link
Copy Source

@DE4DLY commented on GitHub (Mar 5, 2026):

@Arifursdev i don't think you and i would agree on anything....

here. no network, still chatting.

Image
<!-- gh-comment-id:4005997630 --> @DE4DLY commented on GitHub (Mar 5, 2026): @Arifursdev i don't think you and i would agree on anything.... here. no network, still chatting. <img width="2559" height="1438" alt="Image" src="https://github.com/user-attachments/assets/9a45ab5c-0807-4afb-9b05-33f9f12a80b7" />
yindo commented 2026-06-05 14:50:59 -04:00
Author
Owner
Copy Link
Copy Source

@Arifursdev commented on GitHub (Mar 5, 2026):

@DE4DLY you're using ollama so that would be local, but am I wrong in thinking that the default provider ones from anythingllm which are categorized and first category is qwen vision models are local? could you try one of them?

<!-- gh-comment-id:4006041105 --> @Arifursdev commented on GitHub (Mar 5, 2026): @DE4DLY you're using ollama so that would be local, but am I wrong in thinking that the default provider ones from anythingllm which are categorized and first category is qwen vision models are local? could you try one of them?
yindo commented 2026-06-05 14:50:59 -04:00
Author
Owner
Copy Link
Copy Source

@DE4DLY commented on GitHub (Mar 5, 2026):

@Arifursdev i am using the default provider in those screenshots (Which AnythingLLM is running a Ollama Server in the background running qwen3-vl:4b-instruct)

Image
<!-- gh-comment-id:4006064430 --> @DE4DLY commented on GitHub (Mar 5, 2026): @Arifursdev i am using the default provider in those screenshots (Which AnythingLLM is running a Ollama Server in the background running qwen3-vl:4b-instruct) <img width="548" height="133" alt="Image" src="https://github.com/user-attachments/assets/b005d1a9-31ad-422d-820d-4b628e325bfd" />
yindo commented 2026-06-05 14:50:59 -04:00
Author
Owner
Copy Link
Copy Source

@Arifursdev commented on GitHub (Mar 5, 2026):

@DE4DLY correct me If i'm wrong. am I running a downloaded local model or a cloud one for which I assume wouldn't have to download a model?

Image
<!-- gh-comment-id:4006081772 --> @Arifursdev commented on GitHub (Mar 5, 2026): @DE4DLY correct me If i'm wrong. am I running a downloaded local model or a cloud one for which I assume wouldn't have to download a model? <img width="1329" height="732" alt="Image" src="https://github.com/user-attachments/assets/44c6200b-2466-4983-bf7a-268fb92d372d" />
yindo commented 2026-06-05 14:50:59 -04:00
Author
Owner
Copy Link
Copy Source

@DE4DLY commented on GitHub (Mar 5, 2026):

@Arifursdev in this screenshot, you selected the Qwen3 Vision 4B instruct model, just like me. for it to run, it first has to download. when done downloading, it loads it up on your system and then runs fully locally. after that, it won't ever have network activity other than telemetry as explained in the README.md on github, which is totally anonymous and can be turned off on docker (i don't see the option in the settings section on the app, i will bring this up in an issue shortly).

<!-- gh-comment-id:4006108910 --> @DE4DLY commented on GitHub (Mar 5, 2026): @Arifursdev in this screenshot, you selected the Qwen3 Vision 4B instruct model, just like me. for it to run, it first has to download. when done downloading, it loads it up on your system and then runs fully locally. after that, it won't ever have network activity other than telemetry as explained in the README.md on github, which is totally anonymous and can be turned off on docker (i don't see the option in the settings section on the app, i will bring this up in an issue shortly).
yindo commented 2026-06-05 14:51:00 -04:00
Author
Owner
Copy Link
Copy Source

@Arifursdev commented on GitHub (Mar 5, 2026):

@Arifursdev it loads it up on your system and then runs fully locally. after that, it won't ever have network

exactly. try downloading and once its active, try chat it responds instantly. then disconnect the internet and try again, you should find it loading forever in the chat.

<!-- gh-comment-id:4006133988 --> @Arifursdev commented on GitHub (Mar 5, 2026): > [@Arifursdev](https://github.com/Arifursdev) it loads it up on your system and then runs fully locally. after that, it won't ever have network exactly. try downloading and once its active, try chat it responds instantly. then disconnect the internet and try again, you should find it loading forever in the chat.
yindo commented 2026-06-05 14:51:00 -04:00
Author
Owner
Copy Link
Copy Source

@DE4DLY commented on GitHub (Mar 5, 2026):

@Arifursdev i don't think you and i would agree on anything....

here. no network, still chatting.

Image

Like i said here, i have done exactly this, and model stays loaded in memory and responds without delay

<!-- gh-comment-id:4006148976 --> @DE4DLY commented on GitHub (Mar 5, 2026): > [@Arifursdev](https://github.com/Arifursdev) i don't think you and i would agree on anything.... > > here. no network, still chatting. > > <img alt="Image" width="2000" height="1438" src="https://private-user-images.githubusercontent.com/42302714/558804602-9a45ab5c-0807-4afb-9b05-33f9f12a80b7.png?jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.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.hYpsGrlPbbJhQMLGZLg4aDzigbccfRAn62s-hWr1oW8"> Like i said here, i have done exactly this, and model stays loaded in memory and responds without delay
yindo commented 2026-06-05 14:51:00 -04:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Mar 5, 2026):

This is clearly user error, the only "networking" calls for sure are downloading models since we are not about to pre-pack 100GB+ of models into an installer, so you would have to pull them from the cloud. Same applies for embedder/reranker - we pull those as needed too. Same for meeting assistant model - you get the idea.

The only external networking unrelated to runtime, is the telemetry, which can absolutely be disabled in settings -> Privacy & Data -> disable anon telemetry.

Image
<!-- gh-comment-id:4006564381 --> @timothycarambat commented on GitHub (Mar 5, 2026): This is clearly user error, the only "networking" calls for sure are downloading models since we are not about to pre-pack 100GB+ of models into an installer, so you would have to pull them from the cloud. Same applies for embedder/reranker - we pull those as needed too. Same for meeting assistant model - you get the idea. The only external networking unrelated to runtime, is the telemetry, which can absolutely be disabled in settings -> Privacy & Data -> disable anon telemetry. <img width="243" height="213" alt="Image" src="https://github.com/user-attachments/assets/9118de49-9896-41cc-836a-343075af21a7" />
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
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#4934
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?