How to use the ExcaliBrain Obsidian plugin

TL;DR

ExcaliBrain visualizes Obsidian notes by building relationship networks, keeping notes useful through context rather than leaving them as isolated text.

Briefing Cornell Notes

Briefing

ExcaliBrain turns plain Obsidian notes into an interactive, relationship-driven visual map—helping users keep notes useful by placing them in the right context. Instead of tagging images for search (as Excalidraw does), ExcaliBrain builds a “visual exploration” layer over textual writing, making it easier to see what belongs together, what sits above or below, and what’s connected even when the links aren’t obvious.

The plugin’s core value is context. Many notes become irrelevant after school because they’re stored without a meaningful structure. ExcaliBrain addresses that by inferring and/or reading relationships between notes, then rendering them as a graph-like layout in a pop-out window. It relies on two prerequisites—Dataview and Excalidraw—without which it won’t function. Hover Editor can enhance the experience, but it’s optional. Older documentation references Breadcrumbs, but that’s no longer required.

ExcaliBrain’s relationship model is built around four types: parents, children, siblings, and friends. Parents and children form a hierarchy: a parent note is broader in scope (for example, “role playing games”), while a child note is more specific or an implementation (for example, “Dungeons and Dragons” or “D&D”). Friends sit “off to the side”: they share topical overlap without sharing the same parent. Siblings are inferred—notes that would be friends but share the same parent. In practice, ExcaliBrain displays only three relationship types for explicit configuration because siblings are derived automatically from the other relationships.

A key practical point is how relationships get created. At the first level, ExcaliBrain infers structure from links written in notes. If “monitoring” is linked from “observability” using a phrase like “monitoring is a part of observability,” ExcaliBrain treats the linked page as the child and the current page as the parent. Reversing the link flips the hierarchy. It also infers friends when two notes mutually link to each other, producing a dashed connection between them.

The interface supports exploration rather than manual upkeep: users can search within the graph, zoom, drag nodes, click nodes to jump back into Obsidian, and preview notes via hover. It can also reveal “virtual nodes”—linked notes that don’t exist yet—highlighting gaps in a vault. Additional toggles let users show or hide attachments, inferred nodes, page nodes (folders/tags), and aliases, reducing clutter.

ExcaliBrain then scales through deeper usage levels. Level two shifts from inference to explicit relationship keywords (parent/child/friend) embedded in notes, optionally hidden using comment syntax so they don’t clutter reading mode. Level three adds custom relationship keywords tailored to the user’s workflow—such as date published, date, and location—so relationships can reflect chronology or real-world context (e.g., a location note as a parent with people as children). Level four pushes toward semantic relationships, where keywords can express support or opposition (e.g., “reinforces” or “contradicts”), turning the system from hierarchical organization into a tool for critical thinking and ongoing “gardening” of ideas.

Overall, ExcaliBrain is positioned as a personal knowledge management accelerator: once relationships click, it can change how notes are connected, revisited, and tested over time—turning a static vault into a living network of context.

Cornell Notes

ExcaliBrain is an Obsidian plugin that visualizes notes by building relationship networks—so writing stays relevant through context. It can infer relationships from how notes link to each other, using four conceptual types: parents, children, siblings, and friends (with siblings inferred). The plugin requires Dataview and Excalidraw, and it renders an interactive pop-out view where users can search, zoom, and explore connections. As users progress, they can explicitly set relationships with keywords, add custom relationship terms (like dates or locations), and eventually model semantic links such as “reinforces” or “contradicts.” This matters because the system shifts note-taking from storage to ongoing contextual understanding and idea testing.

How does ExcaliBrain decide whether one note is a parent or a child of another?

In the inference mode, ExcaliBrain reads the direction of links written in notes. If “monitoring” is linked from “observability” using wording like “monitoring is a part of observability,” ExcaliBrain treats “observability” as the parent and “monitoring” as the child. Reversing the link—writing the relationship from the “monitoring” note back to “observability”—reverses the hierarchy in the ExcaliBrain view.

What’s the difference between “friends” and “siblings” in ExcaliBrain’s relationship model?

Friends are notes with topical overlap but not the same parent. In the example, “character creation” relates to “Dungeons and Dragons” but also connects to other role-playing and even non-game interests, so it’s a friend rather than a direct child. Siblings are inferred: if two notes would be friends but share the same parent, ExcaliBrain treats them as siblings. The practical takeaway is that siblings are derived from the other relationship structure rather than being explicitly set.

When do two notes become “friends” automatically?

ExcaliBrain infers a friend relationship when two notes mutually link to each other. In the walkthrough, “monitoring” links to “instrumentation” (through the “parts of observability” phrasing), and “instrumentation” links back to “monitoring” with the same kind of mutual relationship. That mutual linking produces a dashed line between the two nodes, indicating they’re both friends and siblings (because they share the same parent).

What does “virtual nodes” mean, and why is it useful?

Virtual nodes are nodes ExcaliBrain can display even when the underlying note doesn’t exist in the vault. If a note is linked but the target page hasn’t been created yet, ExcaliBrain can still show it in the graph. This helps users spot gaps—cases where they’re already referencing something but haven’t actually written the corresponding note.

How does explicit relationship setting (level two) change workflow compared with inference?

With inference, ExcaliBrain derives relationships from how notes link. With explicit setting, users remove the inference-driving links and instead write relationship keywords directly in notes (parent/child/friend). For example, writing “parent: observability” inside “monitoring” makes “observability” the parent even if the earlier linking pattern is gone. Users can also set “friend” relationships explicitly, and siblings can still be inferred from shared parents.

How do custom relationship keywords (level three) expand beyond basic hierarchy?

Custom keywords let users define relationships based on their own fields and patterns. The walkthrough highlights date published/date for chronology and daily-note tracking, and location for CRM-style relationship management—where people can appear as children under a location/country parent. This turns organization into something more meaningful than folders or tags because it reflects how notes relate in the user’s real context.

Review Questions

In inference mode, how would you test whether ExcaliBrain treats note A as the parent of note B versus the child?
Explain how mutual linking can produce a dashed “friend” connection and why siblings may still be inferred at the same time.
What kinds of information could custom keywords like date or location represent in a personal knowledge system, and how would that affect the graph structure?

Key Points

1
ExcaliBrain visualizes Obsidian notes by building relationship networks, keeping notes useful through context rather than leaving them as isolated text.
2
Dataview and Excalidraw are required for ExcaliBrain to work; Hover Editor is optional and Breadcrumbs is no longer necessary.
3
ExcaliBrain’s relationship types include parents, children, siblings, and friends, with siblings inferred from shared parents.
4
Inference mode derives hierarchy from link direction and infers friendships from mutual linking, producing dashed connections in the graph.
5
The pop-out interface supports searching, zooming, dragging, node clicking back into Obsidian, and previewing notes via hover.
6
Virtual nodes reveal links to notes that don’t exist yet, helping users identify missing pages in their vault.
7
Advanced usage adds explicit relationship keywords, custom relationship terms (like date or location), and potentially semantic relationships such as “reinforces” or “contradicts.”

Highlights

ExcaliBrain flips the usual approach: instead of turning visuals into searchable text (Excalidraw), it turns textual notes into a visual relationship map.

Siblings aren’t something users directly set; they’re inferred when two notes share the same parent even if they’re otherwise only “friends.”

Virtual nodes can expose “missing” notes—pages that are referenced but not yet created—so the graph becomes a gap-finding tool.

Semantic relationships (e.g., “reinforces” vs. “contradicts”) are framed as the most advanced path, shifting from hierarchy to idea-level reasoning.

Topics

ExcaliBrain Setup
Relationship Inference
Explicit Relationship Keywords
Virtual Nodes
Semantic Relationships

Mentioned

Obsidian
Excalidraw
Dataview
Hover Editor
Breadcrumbs
ExcaliBrain
Zsolt Vizcain
Nicole van der Hoeven