The knowledge graph, built as you write
Zyler turns [[backlinks]] into a living map of your notes — and gives the AI the tools to read, link, and rewrite across all of them.
Overview
A pile of notes isn't a knowledge base. The value shows up when notes point at each other — when today's meeting links back to the project, the person, and the decision that came before. In Zyler, those connections are first-class, and they double as context for the AI.
Links that don't break
Type [[ and link to any note. Zyler stores the target by a stable noteId, so
renaming a note never breaks a link, and it flags dead links when a target
disappears. Block references go finer still: cite a single block from another
note and watch it stay in sync as the source changes.
A graph you can steer
Every backlink becomes an edge in a force-directed graph view — an Obsidian-style map where nodes are notes and hovering one highlights its neighborhood. You can filter out orphans, tune the physics, and click straight through to a note. It's the fastest way to see what's densely connected and what's drifting alone.
The AI works the graph too
Because links are structured, the assistant can traverse them. Its tools let it:
- search your vault by keyword and pull back the notes that match,
- read the documents you've
@-mentioned or it has found, - create a new note and set it as the editing target, and
- edit blocks directly — insert, update, move, or turn one block into another.
So "summarize what I know about X and link it to the related notes" isn't a prompt into the void — it's the AI reading your graph and writing back into it.
The point of a second brain isn't storage. It's that the connections do work for you later.
Explore it on the Zyler page, where the graph and the assistant run on a sample vault.