[Feature]: Add Source Tracing / Citation functionality to conversation outputs

Description:
Hello! Thank you to the development team for creating such an excellent tool as QwenPaw.

During my daily use, I have a suggestion for improving the user experience: I highly recommend adding a "Source Tracing" or citation feature to the conversation outputs.

When the model processes local files, personal knowledge bases, or provides answers combined with web searches, it would be incredibly helpful if it could annotate "which document was referenced," "which web link it originated from," or specific document snippets after specific paragraphs or sentences, similar to what some mainstream AI products currently do.

Benefits:

It allows users to quickly click and view the original text to verify the accuracy of the provided information.

It helps mitigate the impact of LLM "hallucinations," significantly enhancing the product's credibility and reliability in productivity scenarios.

I hope the team will consider adding this feature to the future roadmap (for example, displaying references via UI cards or superscript badges). Looking forward to seeing QwenPaw get even better. Thank you for your hard work!