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 #4416] [FEAT]: Chunking strategy options #2817

New Issue
Open
opened 2026-02-22 18:31:23 -05:00 by yindo · 3 comments
yindo commented 2026-02-22 18:31:23 -05:00
Owner
Copy Link
Copy Source

Originally created by @alainlizotte on GitHub (Sep 19, 2025).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4416

What would you like to see?

I think the chunking option available right now is the weak point to get better results.

Right now, the chunks size are set in a fixed number of characters.

Would be great to be able to ask for complete sentences or paragraphs only.
Could also use agentic feature to include a complete topic.
The chunks would be limited by the max context lenght (in tokens).

Image
Image
Image
Image
Image

Originally created by @alainlizotte on GitHub (Sep 19, 2025). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4416 ### What would you like to see? I think the chunking option available right now is the weak point to get better results. Right now, the chunks size are set in a fixed number of characters. Would be great to be able to ask for complete sentences or paragraphs only. Could also use agentic feature to include a complete topic. The chunks would be limited by the max context lenght (in tokens). ![Image](https://github.com/user-attachments/assets/73d188ff-3951-4258-a82a-89249e3d45d9) ![Image](https://github.com/user-attachments/assets/c0269d04-9079-4c12-bb47-39f7fac6f4ed) ![Image](https://github.com/user-attachments/assets/59163d2d-e34f-4362-87b2-78b7faf75911) ![Image](https://github.com/user-attachments/assets/141ec11e-da08-48cb-9b3f-f73b0dcdb467) ![Image](https://github.com/user-attachments/assets/c6e7fc7e-7d96-4cfb-bfc4-6e79efabeac3)
yindo added the enhancementfeature request labels 2026-02-22 18:31:23 -05:00
yindo commented 2026-02-22 18:31:24 -05:00
Author
Owner
Copy Link
Copy Source

@TeddyAfro commented on GitHub (Sep 20, 2025):

RAG is one of the main use cases for software such as Anythingllm. With advanced RAG, you would stand out significantly from other providers.

Targeted chunking strategies can significantly improve quality in individual cases:

- MarkdownHeaderTextSplitter
- SemanticChunker
-  RegexTextSplitter
-  SpacyTextSplitter
@TeddyAfro commented on GitHub (Sep 20, 2025): RAG is one of the main use cases for software such as Anythingllm. With advanced RAG, you would stand out significantly from other providers. Targeted chunking strategies can significantly improve quality in individual cases: ``` - MarkdownHeaderTextSplitter - SemanticChunker - RegexTextSplitter - SpacyTextSplitter ```
yindo commented 2026-02-22 18:31:24 -05:00
Author
Owner
Copy Link
Copy Source

@kaidoloor commented on GitHub (Dec 27, 2025):

Assume I have a KM database that I need to query. I would output text files (respecting file size limits) from that KM database (or files) with each paragraph (= line) starting with "[## SOURCE: [...] ]". It would be great if AnythingLLM text chunker would make the chunks starting with the "[## Source:", and ending just before the next "[## Source". Or it would chunk at "[Chunk/split here]". Would that be doable? Or can we make our own chunkers and plug these into AnythingLLM?

@kaidoloor commented on GitHub (Dec 27, 2025): Assume I have a KM database that I need to query. I would output text files (respecting file size limits) from that KM database (or files) with each paragraph (= line) starting with "[## SOURCE: [...] ]". It would be great if AnythingLLM text chunker would make the chunks starting with the "[## Source:", and ending just before the next "[## Source". Or it would chunk at "[Chunk/split here]". Would that be doable? Or can we make our own chunkers and plug these into AnythingLLM?
yindo commented 2026-02-22 18:31:24 -05:00
Author
Owner
Copy Link
Copy Source

@YouareGitHub commented on GitHub (Jan 29, 2026):

yes improved chunking would be great. maybe the ability to pick the chunking method in the interface or even specify a different external chunker as mentioned above.

Something like unstructured.io open source version (https://unstructured.io/) , accepting this might need to be spun up in another docker container so not sure if it could work across desktop app as well.

@YouareGitHub commented on GitHub (Jan 29, 2026): yes improved chunking would be great. maybe the ability to pick the chunking method in the interface or even specify a different external chunker as mentioned above. Something like unstructured.io open source version [(https://unstructured.io/)](https://github.com/Unstructured-IO/unstructured) , accepting this might need to be spun up in another docker container so not sure if it could work across desktop app as well.
yindo changed title from [FEAT]: Chunking strategy options to [GH-ISSUE #4416] [FEAT]: Chunking strategy options 2026-06-05 14:48:42 -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#2817
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?