Context-aware block discovery and one-click integration for long-form writing

While the current "Unlinked Mentions" feature is great for title-matching, it remains too limited for deep, long-form writing (theses, scripts, research). Finding relevant snippets across hundreds of atomic notes often becomes an overwhelming manual task that breaks the writing flow.

The Proposal:

I suggest a discovery mechanism that goes beyond simple title recognition. As I write a sentence or a paragraph, the system should analyze the context and syntax of the draft to surface relevant blocks from my database—regardless of whether the title is mentioned.

Key Features:

• Contextual Block Previews: Instead of just showing the object title, the sidebar should display a preview of the specific block or paragraph that is contextually relevant to my current writing.
• One-Click Integration: A mechanism to instantly "pull" or "insert" that suggested block into the manuscript with a single click (as a text copy, not just an embed).

The Goal:

This would transform Capacities from a "smart storage tool" where notes can get lost, into a high-output production engine. By surfacing relevant data exactly when it’s needed and allowing seamless integration, Capacities becomes a workspace for creating meaningful outputs rather than just managing digital fragments.