AI Assistant should cite notes instead of only generating answers

Hi Capacities team,

I’ve been trying the AI Assistant during the Pro trial, primarily for fiction writing and worldbuilding.

First of all, I really like the direction of Capacities. The object-based model feels much cleaner than traditional note apps, and I can already see the potential.

However, I noticed that the AI Assistant seems optimized for answering questions rather than helping me navigate my own knowledge.

For example, if I ask:

“What is a Horcrux?”

the AI searches my notes and then generates a synthesized explanation.

As a fiction writer, that’s actually less useful than simply showing me the notes it found.

What I would love instead is a mode where the AI behaves more like a research assistant than a chatbot.

For example:

• Show me all relevant notes ranked by relevance.
• Show snippets with the matching text highlighted.
• Explain why each note was selected.
• Optionally generate a synthesis after showing the sources.

This would make it much easier to verify canon, compare old ideas, rediscover forgotten concepts, and avoid hallucinations.

For creative writing, I care much more about recovering my own thoughts than generating new ones.

A workflow like this would be incredible:

Question

What is a Horcrux?

Result

• 📄 Horcrux System
• 📄 Horcrux Extraction
• 📄 Voldemort’s Soulbind Origin
• 📄 Voldemort Character Notes

with highlighted passages, and then an optional “Summarize these notes” button.

I think this would make Capacities one of the strongest tools available for writers, researchers, and anyone maintaining a long-term knowledge base.

Thanks for building such an interesting product!