My notes, articles, and podcasts turned into a connected graph Claude can query, plus a second, separate graph that audits Cloze's help center for contradictions.
graphify reads the source material, extracts concepts and their relationships, de-duplicates, and writes one graph.json that several consumers read.
Brain notes (Projects / Sessions / Stack), articles + news digests, 3,530 podcast summaries.
/graphify --update (I run it), concept + relationship extraction, dedup pass (canonicalize).
brain-graph MCP (7 query tools), LLM wiki: 3,703 files, auto-regen, Obsidian graph view.
Same engine (graphify), two independent corpora, two MCP servers. Kept fully separate by design.
My second brain across everything: notes, reading, listening. Nodes: 21,106. Connections (edges): 21,515. LLM wiki files: 3,703. Served by brain-graph MCP.
A domain audit: Cloze's help center, graphed and fact-checked. Nodes: 2,458. Named communities: 343. Contradictions found: 41. Served by cloze-graph + cloze-audit.
The graph is exposed as a read-only MCP server, so any Claude session can ask it questions. No tokens, no network, just the local graph.
Fetch a concept and everything known about it.
What's directly connected to a concept.
Search the whole graph by topic or term.
How two ideas connect, step by step.
Pull a whole cluster of related notes.
The most-connected hubs. My core themes.
The size and shape of the graph at a glance.
cloze-audit surfaces contradictions, coverage gaps, and redundant articles with sourced quotes.
/graphify --update by hand. It's Max-covered (no metered cost) and keeps a human in the loop, which matters because Brain ingests untrusted web text, an injection surface an unattended agent shouldn't touch.