In GPT and Perplexity we can select some text from a previous prompt or AI answer and then reference it in the next prompt, and it is very useful. How about adding that to cursor chat?
Agreed, this is the one feature that some LLMs have that I really appreciate. This would substantially improve follow-up questions, especially after long or multi-part responses.
Proposed workflow: highlight text in an assistant response, then choose Add to chat as quote from the context menu or use a configurable shortcut. Cursor already supports adding selected code, terminal output, and other IDE context, so applying the same interaction to chat responses would feel natural.
This would let users reference an exact statement without copying it manually or describing its location. A same-conversation implementation alone would already be extremely useful; cross-chat quoting could follow later.
Please consider prioritizing this feature. Thanks.