Mintplex-Labs/anything-llm
1
0
Fork 0
mirror of https://github.com/Mintplex-Labs/anything-llm.git synced 2026-07-21 01:25:28 -04:00
Code Issues 328 Packages Projects Releases 10 Wiki Activity

[GH-ISSUE #4336] [FEAT]: AI Assistant Cannot Access User Profile Bio for Contextual Improvement #2762

New Issue
Open
opened 2026-02-22 18:31:08 -05:00 by yindo · 4 comments
yindo commented 2026-02-22 18:31:08 -05:00
Owner
Copy Link
Copy Source

Originally created by @Wetoshkin on GitHub (Aug 28, 2025).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4336

What would you like to see?

Subject: [Feature Request/Bug Report] AI Assistant Cannot Access User Profile Bio for Contextual Improvement

Description:

I am currently interacting with an AI assistant (powered by LLM technology) that is intended to provide contextualized and personalized responses based on user profile information, specifically the “bio” section. The system prompt indicates that the AI should use this information for better response generation and adaptation.

However, the AI has explicitly stated that it cannot access or read the user’s bio section, even when informed that it has been updated or that it “should” have this capability. This limitation prevents the AI from leveraging potentially valuable contextual data for improved performance.

Expected Behavior:

The AI assistant should be able to access and utilize information present in the user’s “bio” field to:

Personalize responses based on the user’s name, role, or other relevant context.
Adapt its tone and level of detail according to the user’s background.
Better understand the user’s intent and provide more tailored assistance.
Actual Behavior:

The AI assistant consistently reports an inability to access or read the user’s bio. It states: “As a large language model, I do not have the ability to directly access or read information from user profiles or external systems, including ‘bio’ sections.”

Steps to Reproduce (if applicable):

Access a platform where an AI assistant integrates with user profiles, allowing for a “bio” field.
Populate the “bio” field with user-specific information (e.g., a name).
Interact with the AI assistant, prompting it to use the bio for improved responses.
Observe that the AI states it cannot access this information.
Possible Cause/Suggestion for Improvement:

It appears there might be a disconnect between the AI’s intended capabilities (as described in its system prompt) and its actual technical implementation regarding accessing user-specific contextual data from profiles.

Possible solutions could involve:

Explicitly passing the bio information to the AI: Ensure that user profile data (like the bio) is correctly parsed and passed as part of the input context to the LLM during each interaction, possibly via API parameters or pre-processing steps.
Reviewing the AI’s integration with the user profile system: Verify that the necessary permissions and data transfer mechanisms are in place for the AI to read the bio field.
Clarifying AI capabilities in system prompts: If the AI is fundamentally unable to access external profile data, the system prompt should be updated to accurately reflect this limitation, rather than suggesting it can.
Environment:

Platform: AnythingLLM 1.8.4 web
AI Model: DeepSeek

Thank you for your time and for developing this AI assistant. I believe addressing this capability would significantly enhance its usefulness.

Originally created by @Wetoshkin on GitHub (Aug 28, 2025). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4336 ### What would you like to see? Subject: [Feature Request/Bug Report] AI Assistant Cannot Access User Profile Bio for Contextual Improvement Description: I am currently interacting with an AI assistant (powered by LLM technology) that is intended to provide contextualized and personalized responses based on user profile information, specifically the “bio” section. The system prompt indicates that the AI should use this information for better response generation and adaptation. However, the AI has explicitly stated that it cannot access or read the user’s bio section, even when informed that it has been updated or that it “should” have this capability. This limitation prevents the AI from leveraging potentially valuable contextual data for improved performance. Expected Behavior: The AI assistant should be able to access and utilize information present in the user’s “bio” field to: Personalize responses based on the user’s name, role, or other relevant context. Adapt its tone and level of detail according to the user’s background. Better understand the user’s intent and provide more tailored assistance. Actual Behavior: The AI assistant consistently reports an inability to access or read the user’s bio. It states: “As a large language model, I do not have the ability to directly access or read information from user profiles or external systems, including ‘bio’ sections.” Steps to Reproduce (if applicable): Access a platform where an AI assistant integrates with user profiles, allowing for a “bio” field. Populate the “bio” field with user-specific information (e.g., a name). Interact with the AI assistant, prompting it to use the bio for improved responses. Observe that the AI states it cannot access this information. Possible Cause/Suggestion for Improvement: It appears there might be a disconnect between the AI’s intended capabilities (as described in its system prompt) and its actual technical implementation regarding accessing user-specific contextual data from profiles. Possible solutions could involve: Explicitly passing the bio information to the AI: Ensure that user profile data (like the bio) is correctly parsed and passed as part of the input context to the LLM during each interaction, possibly via API parameters or pre-processing steps. Reviewing the AI’s integration with the user profile system: Verify that the necessary permissions and data transfer mechanisms are in place for the AI to read the bio field. Clarifying AI capabilities in system prompts: If the AI is fundamentally unable to access external profile data, the system prompt should be updated to accurately reflect this limitation, rather than suggesting it can. Environment: Platform: AnythingLLM 1.8.4 web AI Model: DeepSeek Thank you for your time and for developing this AI assistant. I believe addressing this capability would significantly enhance its usefulness.
yindo added the enhancementfeature request labels 2026-02-22 18:31:08 -05:00
yindo commented 2026-02-22 18:31:08 -05:00
Author
Owner
Copy Link
Copy Source

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

However, the AI has explicitly stated that it cannot access or read the user’s bio section, even when informed that it has been updated or that it “should” have this capability. This limitation prevents the AI from leveraging potentially valuable contextual data for improved performance.

Understood, this is because of #4160 which should be fixed via #4197

@timothycarambat commented on GitHub (Aug 28, 2025): > However, the AI has explicitly stated that it cannot access or read the user’s bio section, even when informed that it has been updated or that it “should” have this capability. This limitation prevents the AI from leveraging potentially valuable contextual data for improved performance. Understood, this is because of #4160 which should be fixed via #4197
yindo commented 2026-02-22 18:31:08 -05:00
Author
Owner
Copy Link
Copy Source

@Wetoshkin commented on GitHub (Aug 29, 2025):

Hi, @timothycarambat thanks for closing this ticket so quickly and for pointing me to #4160 / #4197.

Just to clarify: when I opened this issue I wasn’t actually talking about the @agent directive. My intention was to describe a limitation in normal conversations with the LLM, outside of the @agent flow.

Specifically, what I noticed is that the assistant does not seem to automatically have access to the user’s bio/profile information during regular chats. I had expected that the LLM would always be able to leverage the bio for personalization (e.g. name, role, tone preferences) even when not using @agent.

So while #4197 looks like a very useful fix for the agent workflow, the gap I’m describing is slightly different:

Bio/context should also be injected into the default LLM conversation flow, not just agent invocations.

Without this, the assistant continues to respond “I cannot access bio information” unless the bio is manually injected into the system prompt.

Would it make sense to consider a separate enhancement where profile fields like bio are automatically passed to the LLM in all conversation modes (not only when @agent is triggered)?

Thanks again for your great work on this project!

@Wetoshkin commented on GitHub (Aug 29, 2025): Hi, @timothycarambat thanks for closing this ticket so quickly and for pointing me to #4160 / #4197. Just to clarify: when I opened this issue I wasn’t actually talking about the @agent directive. My intention was to describe a limitation in normal conversations with the LLM, outside of the @agent flow. Specifically, what I noticed is that the assistant does not seem to automatically have access to the user’s bio/profile information during regular chats. I had expected that the LLM would always be able to leverage the bio for personalization (e.g. name, role, tone preferences) even when not using @agent. So while #4197 looks like a very useful fix for the agent workflow, the gap I’m describing is slightly different: Bio/context should also be injected into the default LLM conversation flow, not just agent invocations. Without this, the assistant continues to respond “I cannot access bio information” unless the bio is manually injected into the system prompt. Would it make sense to consider a separate enhancement where profile fields like bio are automatically passed to the LLM in all conversation modes (not only when @agent is triggered)? Thanks again for your great work on this project!
yindo commented 2026-02-22 18:31:08 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Aug 29, 2025):

Ah then yes, I certainly misunderstood. So yes right now the default system prompt is very simple and to add date or time or bio you need to insert those variables directly.

The reason for bio, specifically, being non default is that you would first have to enable multi user mode for bio to even be available. Because we focus on SLMs and try to prioritize lower end devices we try to be deliberate in what people want in a prompt, so they have to add it themselves.

Now, this would be tedious with many workspaces, so I think there is a larger "default system prompt" idea worth talking about so that when you make a workspace you can get your sensible default, with whatever variables you like to have as default in a system prompt.

@timothycarambat commented on GitHub (Aug 29, 2025): Ah then yes, I certainly misunderstood. So yes right now the default system prompt is very simple and to add date or time or bio you need to insert those variables directly. The reason for bio, specifically, being non default is that you would first have to enable multi user mode for bio to even be available. Because we focus on SLMs and try to prioritize lower end devices we try to be deliberate in what people want in a prompt, so they have to add it themselves. Now, this would be tedious with many workspaces, so I think there is a larger "default system prompt" idea worth talking about so that when you make a workspace you can get your sensible default, with whatever variables you like to have as default in a system prompt.
yindo commented 2026-02-22 18:31:08 -05:00
Author
Owner
Copy Link
Copy Source

@Wetoshkin commented on GitHub (Aug 31, 2025):

Thanks a lot for the clarification! That makes sense regarding multi-user mode and the decision to keep the default system prompt minimal, especially for running on lower-end hardware.

I totally understand the reasoning. My main concern was just about the usability side — for people who do want personalization (like pulling in the user’s bio), it would be great to have a smoother way to do this without having to manually edit every workspace prompt.

I really like your idea of a configurable “default system prompt template” that new workspaces could inherit. That would solve my case perfectly, since I could define once that the bio (and maybe date/time) should always be included, and then every new workspace would have that automatically.

So yes, I’d fully support discussing and prioritizing a “default system prompt” feature. That feels like the cleanest and most flexible approach.

Thanks again for taking the time to explain!

@Wetoshkin commented on GitHub (Aug 31, 2025): Thanks a lot for the clarification! That makes sense regarding multi-user mode and the decision to keep the default system prompt minimal, especially for running on lower-end hardware. I totally understand the reasoning. My main concern was just about the usability side — for people who do want personalization (like pulling in the user’s bio), it would be great to have a smoother way to do this without having to manually edit every workspace prompt. I really like your idea of a configurable “default system prompt template” that new workspaces could inherit. That would solve my case perfectly, since I could define once that the bio (and maybe date/time) should always be included, and then every new workspace would have that automatically. So yes, I’d fully support discussing and prioritizing a “default system prompt” feature. That feels like the cleanest and most flexible approach. Thanks again for taking the time to explain!
yindo changed title from [FEAT]: AI Assistant Cannot Access User Profile Bio for Contextual Improvement to [GH-ISSUE #4336] [FEAT]: AI Assistant Cannot Access User Profile Bio for Contextual Improvement 2026-06-05 14:48:23 -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 enhancement feature request
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#2762
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?