Back to Blog

Meet Karpathy LLM Wiki

An Obsidian plugin that implements Andrej Karpathy's LLM Wiki concept — AI reads your notes and builds a structured wiki with bidirectional links.

The Problem: Notes That Don’t Talk to Each Other

Your Obsidian vault is a goldmine — people, concepts, ideas, connections. But right now they’re just files in folders. Finding what relates to what means searching, tagging, and hoping you remember the thread.

Karpathy’s Insight

In 2024, Andrej Karpathy wrote about a radical idea: treat your notes as raw material, and let an LLM do the architecture work. It reads what you write, pulls out entities and concepts, and weaves them into a structured Wiki — complete with [[bidirectional links]], an auto-generated index, and a chat interface that answers questions from your knowledge.

So you don’t have to be the librarian. No deciding what deserves a page. No maintaining cross-links. No wondering if something is out of date. Drop notes into sources/ and the LLM reads, extracts, writes, links, and even flags contradictions — while you stay in flow.

What This Plugin Does

One command transforms your notes:

  1. Ingest — Drop a note into sources/. The AI extracts people, concepts, and theories, then auto-generates interlinked Wiki pages with bidirectional links.
  2. Query — Ask “what did I write about X?” The answer comes from your own notes, not the internet.
  3. Maintain — The Lint system detects duplicates, dead links, empty pages, and orphans. Smart Fix All resolves them with one click.

Your knowledge grows organically. Every new note finds its place in the existing network. Every contradiction is flagged. Every connection is made.

Built in the Open

This plugin is MIT-licensed and open source. Your data never leaves your Obsidian vault — only sent to your chosen LLM provider during ingest or query. Local models via Ollama keep everything offline.

View on GitHub

What to Read Next