Deconstructing Visual Ideas with Obsidian Excalidraw using Excalidraw Scripts

A new Excalidraw script, “decompose selected elements into new drawing,” automates a core workflow in visual knowledge management: breaking a page built from drawings into reusable “units of idea,” then keeping those units linked so updates propagate upward. Instead of manually copying icons and text into separate note files—and then re-inserting them back into a larger page—the script turns selected elements into a new embedded Excalidraw drawing in one step, preserving the relationship between the atomic idea and the document that uses it.

The script is demonstrated through the idea-decomposition model used in a “book on a page” style layout. A single page functions as a container made of multiple idea units. Each unit can be an embedded Excalidraw drawing with its own internal components (icons, labels, and structure). Those embedded drawings can be reused across different places on the page, and backlinks make navigation practical: clicking an icon inside one embedded drawing can jump to other documents where the same idea appears. The result is a network of connected notes, where the same visual concept—like an icon representing “communication alignment networking collaboration”—serves as a shared building block across multiple higher-level pages.

To make the workflow concrete, the demonstration starts with a simple drawing containing icons such as a light bulb and a hand-washing symbol, representing the idea “hand washing helps stay healthy.” The key problem is reuse: if that thought is useful elsewhere, copying and pasting the elements into a new note file is possible but click-heavy and error-prone. The script replaces that manual process. After installing it via Excalidraw’s “install or update raw scripts” command palette, the user selects the elements and runs the “decompose selected elements into new drawing” action. Excalidraw prompts for a filename, creates a new drawing in the same folder, and embeds it back into the original canvas.

The practical payoff appears when the atomic idea changes. Editing the newly created embedded drawing—such as reshuffling the light bulb and hand icons—automatically updates the parent drawing after a brief moment. That means the “big document” stays consistent with its component ideas, even as those components evolve.

The demonstration also shows multi-level decomposition. A portion of an embedded drawing can be decomposed again into smaller ideas, and backlinks reflect the hierarchy: atomic elements belong to an idea drawing, which belongs to a larger page drawing. As a result, annotations or shapes drawn around a component can appear at the atomic level and then remain visible when returning to the higher-level document. The overall message is that decomposition plus embedding turns visual diagrams into a maintainable, reusable system—like Lego blocks for ideas—while preserving navigability through backlinks and automatic propagation of edits.