1
[Feature Request] Chat: show the answer as it is generated, and show that the model is working
Source: paperless-ngx/paperless-ngx#14059 · opened by @elfensky
Description
Right now, especially when using slow(er) local models, the Chat feature does not show any signs of life until the response comes back.
It would be nice if it could show the model’s thinking as it happens (as an indicator of the connection being alive and processing) stream in the answer as it comes, instead of waiting on the full message.
Today, the chat works, but shows no indicators of progress. No sign if the model received the message, is processing the request, thinking etc...
<img width="512" height="276" alt="image" src=" />
Mocked up locally — the reasoning streams in first, then the answer:
<img width="663" height="690" alt="image" src=" />
Other
_No response_
Right now, especially when using slow(er) local models, the Chat feature does not show any signs of life until the response comes back.
It would be nice if it could show the model’s thinking as it happens (as an indicator of the connection being alive and processing) stream in the answer as it comes, instead of waiting on the full message.
Today, the chat works, but shows no indicators of progress. No sign if the model received the message, is processing the request, thinking etc...
<img width="512" height="276" alt="image" src=" />
Mocked up locally — the reasoning streams in first, then the answer:
<img width="663" height="690" alt="image" src=" />
Other
_No response_
No pledges yet. Be the first to back this.
Comments
Similar requests
[Feature Request] Open the AI document chat as a full page/tab instead of a small overlay window
6 votes · 0 comments
[Feature Request] Chat: show the scope of each question, and optionally show the model's reasoning
1 vote · 0 comments
[Feature Request] AI chat/RAG: cache the embedding model instead of reloading it on every query
1 vote · 0 comments
[Feature Request] Chat overlay: "stick to bottom" only when already at the bottom
3 votes · 0 comments
Add support for Ollama's think: false parameter for AI suggestions
1 vote · 0 comments
No comments yet.