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 #3830] [BUG]: WorkSpace Vector Search API always return empty array #2450

New Issue
Closed
opened 2026-02-22 18:29:43 -05:00 by yindo · 2 comments
yindo commented 2026-02-22 18:29:43 -05:00
Owner
Copy Link
Copy Source

Originally created by @Xheldon on GitHub (May 14, 2025).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/3830

How are you running AnythingLLM?

Docker (local)

What happened?

I send an vector-search request, It's success but the result is {result: []}, and always is. Why?

Are there known steps to reproduce?

  1. I embedding an wiki in a workspace named 'momo'
  2. I Open the /api/docs and in workspace section, I send an chat request, It's work well and return what I want.
  3. I send an vector-search request, It's success but the result is {result: []}, and always is.
  4. Why?

The image:

workspace/momo/chat success:

Image

and the workspace/momo/vector-search always return empty array no matter how I change the query within 200ms:

Image

Note: The Vector DB is the default LanceDB, and the chat model I set is grok-3

Originally created by @Xheldon on GitHub (May 14, 2025). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/3830 ### How are you running AnythingLLM? Docker (local) ### What happened? I send an `vector-search` request, It's success but the result is {result: []}, and always is. Why? ### Are there known steps to reproduce? 1. I embedding an wiki in a workspace named 'momo' 2. I Open the `/api/docs` and in workspace section, I send an `chat` request, It's work well and return what I want. 3. I send an `vector-search` request, It's success but the result is {result: []}, and always is. 4. Why? The image: workspace/momo/chat success: ![Image](https://github.com/user-attachments/assets/ac92b714-5198-4ad5-84bf-ee6aeec4f330) and the workspace/momo/vector-search always return empty array no matter how I change the query within 200ms: ![Image](https://github.com/user-attachments/assets/887b0583-42aa-493e-ac84-f5f82f81e21f) Note: The Vector DB is the default LanceDB, and the chat model I set is grok-3
yindo added the possible bug label 2026-02-22 18:29:43 -05:00
yindo closed this issue 2026-02-22 18:29:43 -05:00
yindo commented 2026-02-22 18:29:44 -05:00
Author
Owner
Copy Link
Copy Source

@Xheldon commented on GitHub (May 14, 2025):

Sorry，I resolve it，It's just the params problem: I set topN: 4, scoreThreshold: 0.2, will work, but 0.2 is to low...It's the embedding model problem or my wiki?

@Xheldon commented on GitHub (May 14, 2025): Sorry，I resolve it，It's just the params problem: I set topN: 4, scoreThreshold: 0.2, will work, but 0.2 is to low...It's the embedding model problem or my wiki?
yindo commented 2026-02-22 18:29:44 -05:00
Author
Owner
Copy Link
Copy Source

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

Its worth noting the sources in the chat request is empty - this is what are the citations or the vector results of the query as well. If the embedding result is too low this can very well be the embedding model, but it depends on the use case and language. There are embedding models specifically for code, other languages, etc.

The default embedder is kind of ok at code, but its primary use is english/latin-based text. Also we have reranking which will help with low relevancy scores from vanilla semantic search

@timothycarambat commented on GitHub (May 15, 2025): Its worth noting the `sources` in the chat request is empty - this is what are the `citations` or the vector results of the query as well. If the embedding result is too low this can very well be the embedding model, but it depends on the use case and language. There are embedding models specifically for code, other languages, etc. The default embedder is *kind of ok* at code, but its primary use is english/latin-based text. Also we have [reranking](https://docs.anythingllm.com/llm-not-using-my-docs#vector-database-settings--search-preference) which will help with low relevancy scores from vanilla semantic search
yindo changed title from [BUG]: WorkSpace Vector Search API always return empty array to [GH-ISSUE #3830] [BUG]: WorkSpace Vector Search API always return empty array 2026-06-05 14:46:36 -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#2450
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?