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 #1661] [FEAT]: Pre-Processing User Queries with LLM for Improved Similarity Search #1081

New Issue
Closed
opened 2026-02-22 18:23:02 -05:00 by yindo · 3 comments
yindo commented 2026-02-22 18:23:02 -05:00
Owner
Copy Link
Copy Source

Originally created by @khalilxg on GitHub (Jun 12, 2024).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/1661

What would you like to see?

Add a feature where user queries are first processed by an LLM to better match the indexed data. This processed query is then sent to the vector database for improved search results.

Workflow:

User submits a query.
Query is processed by an LLM to align with indexed data.
Processed query is sent to the vector database.
Search results are returned to the user.
Benefits:

Enhanced search accuracy.
More relevant results.
Improved user experience.
Example:

Original Query: "How to fix login issues?"
LLM Processed Query: "Resolving authentication errors in user access"
Thank you for considering this request.

Other benefits: extract matadata from query that suits db || translate user query into general language of embeded-models exemple: translate chinese query to be translated by LLM into english to match db data that is in english

Originally created by @khalilxg on GitHub (Jun 12, 2024). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/1661 ### What would you like to see? Add a feature where user queries are first processed by an LLM to better match the indexed data. This processed query is then sent to the vector database for improved search results. Workflow: User submits a query. Query is processed by an LLM to align with indexed data. Processed query is sent to the vector database. Search results are returned to the user. Benefits: Enhanced search accuracy. More relevant results. Improved user experience. Example: Original Query: "How to fix login issues?" LLM Processed Query: "Resolving authentication errors in user access" Thank you for considering this request. Other benefits: extract matadata from query that suits db || translate user query into general language of embeded-models exemple: translate chinese query to be translated by LLM into english to match db data that is in english
yindo added the enhancementfeature request labels 2026-02-22 18:23:02 -05:00
yindo closed this issue 2026-02-22 18:23:02 -05:00
yindo commented 2026-02-22 18:23:02 -05:00
Author
Owner
Copy Link
Copy Source

@khalilxg commented on GitHub (Jun 17, 2024):

Anyone worked on this ?

@khalilxg commented on GitHub (Jun 17, 2024): Anyone worked on this ?
yindo commented 2026-02-22 18:23:02 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Jun 17, 2024):

Closing this as not planned. Any pre-processing steps would be very use-case dependent and results of using an LLM to "expand" a query for better RAG results are typically better handled via implementation of knowledge graph support and even reranking models. These results even then tend to be highly dependent on the documents used and increase latency by a lot with minimal improvements to the end result.

however, its worth mentioning that this really I think would be better handled via a middleware/plugin system that can run in AnythingLLM that a user can inject. This way, the tool can still support the workflow you intend to use but does not have to be maintained by the core-team since any specific pre/post processing to a query would likely be unused by some greater population of users.

@timothycarambat commented on GitHub (Jun 17, 2024): Closing this as not planned. Any pre-processing steps would be very use-case dependent and results of using an LLM to "expand" a query for better RAG results are typically better handled via implementation of knowledge graph support and even reranking models. These results even then tend to be highly dependent on the documents used and increase latency by a lot with minimal improvements to the end result. _however_, its worth mentioning that this really I think would be better handled via a middleware/plugin system that can run in AnythingLLM that a user can inject. This way, the tool can still support the workflow you intend to use but does not have to be maintained by the core-team since any specific pre/post processing to a query would likely be unused by some greater population of users.
yindo commented 2026-02-22 18:23:02 -05:00
Author
Owner
Copy Link
Copy Source

@khalilxg commented on GitHub (Jun 17, 2024):

Thank you!

@khalilxg commented on GitHub (Jun 17, 2024): Thank you!
yindo changed title from [FEAT]: Pre-Processing User Queries with LLM for Improved Similarity Search to [GH-ISSUE #1661] [FEAT]: Pre-Processing User Queries with LLM for Improved Similarity Search 2026-06-05 14:38:53 -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#1081
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?