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 #5242] [BUG]: Automatic Mode toggle missing in workspace settings on Docker self-hosted install with Generic OpenAI + native tool calling enabled #4984

New Issue
Closed
opened 2026-06-05 14:51:17 -04:00 by yindo · 7 comments
yindo commented 2026-06-05 14:51:17 -04:00
Owner
Copy Link
Copy Source

Originally created by @raucodes on GitHub (Mar 20, 2026).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5242

How are you running AnythingLLM?

Docker (local)

What happened?

I’m running AnythingLLM self-hosted via Docker and cannot enable Automatic Mode for an existing workspace, even though the current documentation suggests this should be possible from Workspace Settings → Chat Settings.

According to the docs, versions newer than v1.11.1 should create new workspaces with Automatic Mode enabled by default, while older/existing workspaces should be manually updatable to Automatic Mode from the workspace settings. In my case, that option is not available in the workspace UI, and the prompt still shows @, meaning I still have to manually invoke @agent for web search / agentic behavior.

My setup uses Generic OpenAI with native tool calling intentionally enabled through the environment variable:
PROVIDER_SUPPORTS_NATIVE_TOOL_CALLING="generic-openai"

So this does not appear to be just a missing native-tool-calling flag. The provider is explicitly allowed for native tool calling, but the workspace still cannot be switched to Automatic Mode.

This is confusing because the documentation makes it sound like:

  1. Automatic Mode is a general workspace chat mode
  2. Existing workspaces can be manually updated
  3. Generic OpenAI should work for native tool calling when the env flag is set

What I expected:

  • A visible Automatic Mode option in the workspace Chat Settings
  • Or automatic migration / availability for the existing workspace after upgrading
  • Or at minimum documentation that this is currently not supported in Docker / self-hosted / Generic OpenAI / old workspaces

What actually happens:

  • No Automatic Mode toggle is shown in the workspace settings
  • The prompt still displays @
  • I must still explicitly use @agent for web search and tool usage

Please confirm whether this is:

  • a bug in Docker/self-hosted,
  • a migration issue for older workspaces,
  • a limitation specific to Generic OpenAI,
  • or a docs mismatch.

Environment

  • Deployment: Docker (self-hosted)
  • Version: 1.11.2
  • Provider: Generic OpenAI
  • Native tool calling env enabled: yes
  • Workspace created before/after upgrade: at 1.11.1

Are there known steps to reproduce?

No response

Originally created by @raucodes on GitHub (Mar 20, 2026). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5242 ### How are you running AnythingLLM? Docker (local) ### What happened? I’m running AnythingLLM self-hosted via Docker and cannot enable **Automatic Mode** for an existing workspace, even though the current documentation suggests this should be possible from **Workspace Settings → Chat Settings**. According to the docs, versions newer than **v1.11.1** should create new workspaces with Automatic Mode enabled by default, while older/existing workspaces should be manually updatable to Automatic Mode from the workspace settings. In my case, that option is not available in the workspace UI, and the prompt still shows `@`, meaning I still have to manually invoke `@agent` for web search / agentic behavior. My setup uses **Generic OpenAI** with native tool calling intentionally enabled through the environment variable: `PROVIDER_SUPPORTS_NATIVE_TOOL_CALLING="generic-openai"` So this does not appear to be just a missing native-tool-calling flag. The provider is explicitly allowed for native tool calling, but the workspace still cannot be switched to Automatic Mode. This is confusing because the documentation makes it sound like: 1. Automatic Mode is a general workspace chat mode 2. Existing workspaces can be manually updated 3. Generic OpenAI should work for native tool calling when the env flag is set What I expected: * A visible **Automatic Mode** option in the workspace Chat Settings * Or automatic migration / availability for the existing workspace after upgrading * Or at minimum documentation that this is currently not supported in Docker / self-hosted / Generic OpenAI / old workspaces What actually happens: * No Automatic Mode toggle is shown in the workspace settings * The prompt still displays `@` * I must still explicitly use `@agent` for web search and tool usage Please confirm whether this is: * a bug in Docker/self-hosted, * a migration issue for older workspaces, * a limitation specific to Generic OpenAI, * or a docs mismatch. **Environment** * Deployment: Docker (self-hosted) * Version: 1.11.2 * Provider: Generic OpenAI * Native tool calling env enabled: yes * Workspace created before/after upgrade: at 1.11.1 ### Are there known steps to reproduce? _No response_
yindo added the possible bug label 2026-06-05 14:51:17 -04:00
yindo closed this issue 2026-06-05 14:51:17 -04:00
yindo commented 2026-06-05 14:51:18 -04:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Mar 20, 2026):

According to the docs, versions newer than v1.11.1 should create new workspaces with Automatic Mode enabled by default, while older/existing workspaces should be manually updatable to Automatic Mode from the workspace settings. In my case, that option is not available in the workspace UI, and the prompt still shows @, meaning I still have to manually invoke @agent for web search / agentic behavior.

This is an error, I accidentally merged this and have yet to actually force this behavior as we wanted to do a slow rollout instead (#5144) and so right now all new workspaces are actually in chat mode still - you need to manually change them to auto.

So this is a docs mismatch and jumped the gun - just patch in docs with d9196b6946f1274b1b9266a03796f72ace8619ac

<!-- gh-comment-id:4101309420 --> @timothycarambat commented on GitHub (Mar 20, 2026): > According to the docs, versions newer than v1.11.1 should create new workspaces with Automatic Mode enabled by default, while older/existing workspaces should be manually updatable to Automatic Mode from the workspace settings. In my case, that option is not available in the workspace UI, and the prompt still shows @, meaning I still have to manually invoke @agent for web search / agentic behavior. This is an error, I accidentally merged this and have yet to actually force this behavior as we wanted to do a slow rollout instead (#5144) and so right now all new workspaces are actually in `chat` mode still - you need to manually change them to auto. So this is a docs mismatch and jumped the gun - just patch in docs with d9196b6946f1274b1b9266a03796f72ace8619ac
yindo commented 2026-06-05 14:51:18 -04:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Mar 20, 2026):

However, the toggle should 100% be present - it is just a new option now on any workspace. It has some provider gaps, but this can be resolved as stated in the docs via the manual ENV key setting for now
https://docs.anythingllm.com/configuration#native-tool-calling-for-llm-providers

<!-- gh-comment-id:4101313902 --> @timothycarambat commented on GitHub (Mar 20, 2026): However, the toggle should 100% be present - it is just a new option now on any workspace. It has some provider gaps, but this can be resolved as stated in the docs via the manual ENV key setting for now https://docs.anythingllm.com/configuration#native-tool-calling-for-llm-providers
yindo commented 2026-06-05 14:51:18 -04:00
Author
Owner
Copy Link
Copy Source

@raucodes commented on GitHub (Mar 21, 2026):

environment: DATABASE_URL: postgresql://anythingllm:foo@bar:5432/anythingllm STORAGE_DIR: /app/server/storage PROVIDER_SUPPORTS_NATIVE_TOOL_CALLING: bedrock,generic-openai,groq,lemonade,litellm,local-ai,openrouter

Image

like I said it's not there in the PG docker container.

<!-- gh-comment-id:4102980032 --> @raucodes commented on GitHub (Mar 21, 2026): ` environment: DATABASE_URL: postgresql://anythingllm:foo@bar:5432/anythingllm STORAGE_DIR: /app/server/storage PROVIDER_SUPPORTS_NATIVE_TOOL_CALLING: bedrock,generic-openai,groq,lemonade,litellm,local-ai,openrouter` <img width="1630" height="405" alt="Image" src="https://github.com/user-attachments/assets/f0d25751-e7c6-4e67-93dd-aae7f0950625" /> like I said it's not there in the PG docker container.
yindo commented 2026-06-05 14:51:18 -04:00
Author
Owner
Copy Link
Copy Source

@raucodes commented on GitHub (Mar 21, 2026):

Test container with latest tag not pg

Image
<!-- gh-comment-id:4103060665 --> @raucodes commented on GitHub (Mar 21, 2026): Test container with latest tag not pg <img width="1630" height="652" alt="Image" src="https://github.com/user-attachments/assets/364b734d-8f2b-49de-8e45-09ec3b39c346" />
yindo commented 2026-06-05 14:51:18 -04:00
Author
Owner
Copy Link
Copy Source

@raucodes commented on GitHub (Mar 21, 2026):

Manually setting chatMode='automatic' in the database does not enable Auto Mode in the UI or chat behavior.
This suggests the database field is not the actual source of truth, and Auto Mode availability is being gated elsewhere at runtime.
Since latest shows Auto while :pg does not, this still looks like an image/runtime-specific issue rather than a migration or stale-data problem.

<!-- gh-comment-id:4103288634 --> @raucodes commented on GitHub (Mar 21, 2026): Manually setting chatMode='automatic' in the database does not enable Auto Mode in the UI or chat behavior. This suggests the database field is not the actual source of truth, and Auto Mode availability is being gated elsewhere at runtime. Since latest shows Auto while :pg does not, this still looks like an image/runtime-specific issue rather than a migration or stale-data problem.
yindo commented 2026-06-05 14:51:18 -04:00
Author
Owner
Copy Link
Copy Source

@raucodes commented on GitHub (Mar 22, 2026):

confirmed this is not a general PostgreSQL limitation.
I merged current master into the pg branch locally, built it, and Automatic Mode is now present while Prisma still runs against PostgreSQL.
So the issue appears to be that the published :pg image / pg release path is lagging behind the relevant master changes, not that PostgreSQL itself prevents Automatic Mode.

<!-- gh-comment-id:4105790789 --> @raucodes commented on GitHub (Mar 22, 2026): confirmed this is not a general PostgreSQL limitation. I merged current master into the pg branch locally, built it, and Automatic Mode is now present while Prisma still runs against PostgreSQL. So the issue appears to be that the published :pg image / pg release path is lagging behind the relevant master changes, not that PostgreSQL itself prevents Automatic Mode.
yindo commented 2026-06-05 14:51:18 -04:00
Author
Owner
Copy Link
Copy Source

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

1.11.2 was release on March 18 at 9AM PT
Image

Automatic mode was merged in March 18 at 12PM PT
Image

1.11.2 would not have this change since it was merged after the tag was published - that is why it does not exist in pg or pg-1.11.2. That change is part of >1.11.2, not 1.11.2.

pg ≠ latest either: https://docs.anythingllm.com/installation-docker/available-images#pg

Which is why the base does not have it either, since it is synced to latest published version, no the mainline latest.

<!-- gh-comment-id:4112063564 --> @timothycarambat commented on GitHub (Mar 23, 2026): 1.11.2 was release on **March 18 at 9AM PT** <img width="956" height="144" alt="Image" src="https://github.com/user-attachments/assets/646d7022-03d3-4bd0-9b91-272b27c09817" /> Automatic mode was merged in **March 18 at 12PM PT** <img width="942" height="281" alt="Image" src="https://github.com/user-attachments/assets/d9dc85fb-addd-44f3-9bd2-2d2f0f81dcf0" /> 1.11.2 **would not** have this change since it was merged after the tag was published - that is why it does not exist in pg or pg-1.11.2. That change is part of >1.11.2, not 1.11.2. pg ≠ latest either: https://docs.anythingllm.com/installation-docker/available-images#pg Which is why the base does not have it either, since it is synced to latest published version, no the mainline latest.
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#4984
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?