[GH-ISSUE #4558] [FEAT]: Document import function #2897

Closed
opened 2026-02-22 18:31:43 -05:00 by yindo · 3 comments
Owner

Originally created by @BenjaminMuslic on GitHub (Oct 17, 2025).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4558

What would you like to see?

The paper-clip upload currently defaults to RAG (embedding + retrieval). For some use cases, we need the entire document text present in the model’s context for deterministic reasoning. Copy-pasting is clunky and error-prone, and retrieval can miss key passages.

Proposed Solution

Add a second ingestion mode for uploads: Full-Context Import. This mode places the uploaded file’s plain text directly into the next model call’s prompt (no embedding).

Originally created by @BenjaminMuslic on GitHub (Oct 17, 2025). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4558 ### What would you like to see? The paper-clip upload currently defaults to RAG (embedding + retrieval). For some use cases, we need the entire document text present in the model’s context for deterministic reasoning. Copy-pasting is clunky and error-prone, and retrieval can miss key passages. Proposed Solution Add a second ingestion mode for uploads: Full-Context Import. This mode places the uploaded file’s plain text directly into the next model call’s prompt (no embedding).
yindo added the enhancementfeature request labels 2026-02-22 18:31:43 -05:00
yindo closed this issue 2026-02-22 18:31:43 -05:00
Author
Owner

@timothycarambat commented on GitHub (Oct 17, 2025):

The paper-clip upload currently defaults to RAG (embedding + retrieval). For some use cases, we need the entire document text present in the model’s context for deterministic reasoning. Copy-pasting is clunky and error-prone, and retrieval can miss key passages.

This is not true as of 1.8.5: https://github.com/Mintplex-Labs/anything-llm/releases/tag/v1.8.5

You are on an old version as what you describe is exactly how the system works now. We even prompt when your document is too large for context to then optionally embed.

@timothycarambat commented on GitHub (Oct 17, 2025): > The paper-clip upload currently defaults to RAG (embedding + retrieval). For some use cases, we need the entire document text present in the model’s context for deterministic reasoning. Copy-pasting is clunky and error-prone, and retrieval can miss key passages. This is not true as of 1.8.5: https://github.com/Mintplex-Labs/anything-llm/releases/tag/v1.8.5 You are on an old version as what you describe is exactly how the system works now. We even prompt when your document is too large for context to then _optionally_ embed.
Author
Owner

@BenjaminMuslic commented on GitHub (Oct 17, 2025):

Oh perfect! Thanks for your quick response Tim! We will update our version then.

@BenjaminMuslic commented on GitHub (Oct 17, 2025): Oh perfect! Thanks for your quick response Tim! We will update our version then.
Author
Owner

@timothycarambat commented on GitHub (Oct 17, 2025):

Let me know if you don't see those changes (not sure how you are running AnythingLLM) but hopefully this improves your experience. It was a lot of work to do and was very much needed for exactly why you made this issue.

@timothycarambat commented on GitHub (Oct 17, 2025): Let me know if you don't see those changes (not sure how you are running AnythingLLM) but hopefully this improves your experience. It was a lot of work to do and was **very much needed** for exactly why you made this issue.
yindo changed title from [FEAT]: Document import function to [GH-ISSUE #4558] [FEAT]: Document import function 2026-06-05 14:49:08 -04:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Mintplex-Labs/anything-llm#2897