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 #4151] [BUG]: [Bug Report] VPN API Connection Failure & Gemini Model List Not Updated #2639

New Issue
Closed
opened 2026-02-22 18:30:33 -05:00 by yindo · 4 comments
yindo commented 2026-02-22 18:30:33 -05:00
Owner
Copy Link
Copy Source

Originally created by @momusticks on GitHub (Jul 15, 2025).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4151

How are you running AnythingLLM?

AnythingLLM desktop app

What happened?

[Body:]

Dear AnythingLLM Development Team,

Hello!

First and foremost, thank you so much for developing such an excellent personal knowledge base tool as AnythingLLM. I'm a big fan of its concept and functionality, and it plays an important role in my work and studies.

During my recent use, I have encountered two issues that I hope you can address:

Issue 1: API Services Fail to Connect in a VPN Environment

When my VPN is active, any attempt to use API services from Google Gemini or OpenRouter results in the error message: "Could not respond to message. 400 status code (no body)". This prevents me from interacting with the models.

To troubleshoot this, I tested another similar knowledge base software, Cherry Studio, under the exact same network conditions, and it was able to connect to and use the aforementioned APIs without any issues. Therefore, I suspect this might be related to how AnythingLLM handles network requests within a VPN environment.

Steps to Reproduce:

Activate a VPN service.

In AnythingLLM, configure the API key for Google Gemini or OpenRouter and select the corresponding AI model.

Attempt to send a message to the configured model.

The system returns the "Could not respond to message. 400 status code (no body)" error.

Expected Behavior:
AnythingLLM should be able to connect to and use API services from platforms like Gemini and OpenRouter seamlessly within a VPN environment, just as it does under normal network conditions.

Issue 2: Latest Model Not Displayed in the "Gemini" LLM Provider List

I have noticed that in the "LLM Provider" settings, the list of available models for "Gemini" does not include the latest "Gemini 2.5 Pro" model. This prevents users who wish to experience the capabilities of the newest and most powerful models from doing so.

Steps to Reproduce:

Go to "Settings" -> "LLM Providers".

Select "Gemini".

View the "Chat Model" dropdown list.

Expected Behavior:
The "Chat Model" list should include "Gemini 2.5 Pro" as an option for users to select and use.

I fully understand that encountering bugs is an inevitable part of the software development process. I am reporting these issues to help make AnythingLLM even more robust and powerful.

Thank you again for your hard work and dedication! I look forward to these issues being resolved soon. Please let me know if you need any more detailed information from my end, such as log files or more specific environment configurations.

Best regards,

A loyal user of AnythingLLM

Are there known steps to reproduce?

Image
Originally created by @momusticks on GitHub (Jul 15, 2025). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4151 ### How are you running AnythingLLM? AnythingLLM desktop app ### What happened? [Body:] Dear AnythingLLM Development Team, Hello! First and foremost, thank you so much for developing such an excellent personal knowledge base tool as AnythingLLM. I'm a big fan of its concept and functionality, and it plays an important role in my work and studies. During my recent use, I have encountered two issues that I hope you can address: Issue 1: API Services Fail to Connect in a VPN Environment When my VPN is active, any attempt to use API services from Google Gemini or OpenRouter results in the error message: "Could not respond to message. 400 status code (no body)". This prevents me from interacting with the models. To troubleshoot this, I tested another similar knowledge base software, Cherry Studio, under the exact same network conditions, and it was able to connect to and use the aforementioned APIs without any issues. Therefore, I suspect this might be related to how AnythingLLM handles network requests within a VPN environment. Steps to Reproduce: Activate a VPN service. In AnythingLLM, configure the API key for Google Gemini or OpenRouter and select the corresponding AI model. Attempt to send a message to the configured model. The system returns the "Could not respond to message. 400 status code (no body)" error. Expected Behavior: AnythingLLM should be able to connect to and use API services from platforms like Gemini and OpenRouter seamlessly within a VPN environment, just as it does under normal network conditions. Issue 2: Latest Model Not Displayed in the "Gemini" LLM Provider List I have noticed that in the "LLM Provider" settings, the list of available models for "Gemini" does not include the latest "Gemini 2.5 Pro" model. This prevents users who wish to experience the capabilities of the newest and most powerful models from doing so. Steps to Reproduce: Go to "Settings" -> "LLM Providers". Select "Gemini". View the "Chat Model" dropdown list. Expected Behavior: The "Chat Model" list should include "Gemini 2.5 Pro" as an option for users to select and use. I fully understand that encountering bugs is an inevitable part of the software development process. I am reporting these issues to help make AnythingLLM even more robust and powerful. Thank you again for your hard work and dedication! I look forward to these issues being resolved soon. Please let me know if you need any more detailed information from my end, such as log files or more specific environment configurations. Best regards, A loyal user of AnythingLLM ### Are there known steps to reproduce? <img width="1710" height="1026" alt="Image" src="https://github.com/user-attachments/assets/2f4dc067-d142-4374-971f-f8f886be0aa1" />
yindo added the possible bug label 2026-02-22 18:30:33 -05:00
yindo closed this issue 2026-02-22 18:30:33 -05:00
yindo commented 2026-02-22 18:30:34 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Jul 15, 2025):

Interesting since we simply should just be relaying the request using your client-side connection - so VPN or not should not matter too much. I know there is a dropped issue of #3589 that is related, because it is a bit out of scope for us.

That being said, I dont use a VPN so replication of an issue like this has been troublesome to try to replicate or what even a sustainable solution would be that makes sense for us to own, as opposed to the client setting up themselves.

@timothycarambat commented on GitHub (Jul 15, 2025): Interesting since we simply should just be relaying the request using your client-side connection - so VPN or not should not matter too much. I know there is a dropped issue of #3589 that is related, because it is a bit out of scope for us. That being said, I dont use a VPN so replication of an issue like this has been troublesome to try to replicate or what even a sustainable solution would be that makes sense for us to own, as opposed to the client setting up themselves.
yindo commented 2026-02-22 18:30:34 -05:00
Author
Owner
Copy Link
Copy Source

@momusticks commented on GitHub (Jul 15, 2025):

Interesting since we simply should just be relaying the request using your client-side connection - so VPN or not should not matter too much. I know there is a dropped issue of #3589 that is related, because it is a bit out of scope for us.

That being said, I dont use a VPN so replication of an issue like this has been troublesome to try to replicate or what even a sustainable solution would be that makes sense for us to own, as opposed to the client setting up themselves.

Thank you so much for your prompt reply and for sharing your insights on the VPN connection issue.

I understand your perspective on network configuration issues like VPNs being complex and potentially outside the core scope of the application. Your explanation regarding issue #3589 makes it clear why handling such specific user-side network setups can be challenging. I appreciate the transparency.

I would also like to gently follow up on the second issue I raised in my initial report, which might have been overlooked. The list of available models for the 'Gemini' LLM provider in the settings doesn't currently include Google's latest models. The latest stable version is Gemini 2.5 Pro. This model offers significant improvements, and enabling it would allow users to leverage its enhanced capabilities directly within AnythingLLM.

On a related note, I was wondering if you might consider adding support for another promising platform: Moonshot AI. Their latest model, Kimi K2, has recently generated a lot of excitement.Released in July 2025, Kimi K2 is a powerful Mixture-of-Experts model that excels at tasks requiring advanced reasoning, coding, and especially "agentic" abilities (the model can act on its own to complete complex tasks).I believe that supporting Kimi would be a fantastic addition, attracting even more users who are keen on exploring cutting-edge AI.

Thank you again for creating such a wonderful tool and for your continuous hard work. I truly appreciate your dedication to the project and look forward to its future developments.

@momusticks commented on GitHub (Jul 15, 2025): > Interesting since we simply should just be relaying the request using your client-side connection - so VPN or not should not matter too much. I know there is a dropped issue of [#3589](https://github.com/Mintplex-Labs/anything-llm/issues/3589) that is related, because it is a bit out of scope for us. > > That being said, I dont use a VPN so replication of an issue like this has been troublesome to try to replicate or what even a sustainable solution would be that makes sense for us to own, as opposed to the client setting up themselves. Thank you so much for your prompt reply and for sharing your insights on the VPN connection issue. I understand your perspective on network configuration issues like VPNs being complex and potentially outside the core scope of the application. Your explanation regarding issue #3589 makes it clear why handling such specific user-side network setups can be challenging. I appreciate the transparency. I would also like to gently follow up on the second issue I raised in my initial report, which might have been overlooked. The list of available models for the 'Gemini' LLM provider in the settings doesn't currently include Google's latest models. The latest stable version is Gemini 2.5 Pro. This model offers significant improvements, and enabling it would allow users to leverage its enhanced capabilities directly within AnythingLLM. On a related note, I was wondering if you might consider adding support for another promising platform: Moonshot AI. Their latest model, Kimi K2, has recently generated a lot of excitement.Released in July 2025, Kimi K2 is a powerful Mixture-of-Experts model that excels at tasks requiring advanced reasoning, coding, and especially "agentic" abilities (the model can act on its own to complete complex tasks).I believe that supporting Kimi would be a fantastic addition, attracting even more users who are keen on exploring cutting-edge AI. Thank you again for creating such a wonderful tool and for your continuous hard work. I truly appreciate your dedication to the project and look forward to its future developments.
yindo commented 2026-02-22 18:30:34 -05:00
Author
Owner
Copy Link
Copy Source

@momusticks commented on GitHub (Jul 15, 2025):

Interesting since we simply should just be relaying the request using your client-side connection - so VPN or not should not matter too much. I know there is a dropped issue of #3589 that is related, because it is a bit out of scope for us.

That being said, I dont use a VPN so replication of an issue like this has been troublesome to try to replicate or what even a sustainable solution would be that makes sense for us to own, as opposed to the client setting up themselves.

I sincerely apologize — upon further investigation, I’ve realized that both issues mentioned above were caused by my expired API key.😭

@momusticks commented on GitHub (Jul 15, 2025): > Interesting since we simply should just be relaying the request using your client-side connection - so VPN or not should not matter too much. I know there is a dropped issue of [#3589](https://github.com/Mintplex-Labs/anything-llm/issues/3589) that is related, because it is a bit out of scope for us. > > That being said, I dont use a VPN so replication of an issue like this has been troublesome to try to replicate or what even a sustainable solution would be that makes sense for us to own, as opposed to the client setting up themselves. I sincerely apologize — upon further investigation, I’ve realized that both issues mentioned above were caused by my expired API key.😭
yindo commented 2026-02-22 18:30:34 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Jul 15, 2025):

@momusticks haha, it happens.

WRT to moonshot -> https://github.com/Mintplex-Labs/anything-llm/issues/4155
made an issue for it so it is tracked. You should be able to use the Generic OpenAI connector in the meantime and that should work

@timothycarambat commented on GitHub (Jul 15, 2025): @momusticks haha, it happens. WRT to moonshot -> https://github.com/Mintplex-Labs/anything-llm/issues/4155 made an issue for it so it is tracked. You _should_ be able to use the Generic OpenAI connector in the meantime and that should work
yindo changed title from [BUG]: [Bug Report] VPN API Connection Failure & Gemini Model List Not Updated to [GH-ISSUE #4151] [BUG]: [Bug Report] VPN API Connection Failure & Gemini Model List Not Updated 2026-06-05 14:47:41 -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#2639
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?