Feature Request: Allow editing/removing attachments from conversation messages

[allozaur]

Feature Description

Enable users to delete, replace, or add media to an existing conversation and re-execute OCR/text processing using the same prompt, without restarting the chat.

Motivation

Currently, users cannot modify media attachments (images, PDFs, audio) in an existing conversation—they must start a new chat to replace media¹ or re-run processes like OCR². This is inefficient for workflows requiring repeated use of the same prompt (e.g., Nanonet-ocr for Qwen3 VL models), as users must:

• Re-enter the same prompt manually.
• Lose conversation context.
• Wasteful for tasks where media is updated incrementally (e.g., refining OCR results with new images).

Footnotes

1. https://github.com/ggml-org/llama.cpp/pull/14839#issuecomment-3304216662 ↩

2. https://www.reddit.com/r/LocalLLaMA/comments/1ooa342/comment/nn2xqfv/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_button ↩