llm-wiki

A reference implementation of Andrej Karpathy's LLM Wiki pattern — a compounding personal knowledge base where the LLM maintains structured, interlinked markdown instead of re-deriving everything from raw chunks on every question.

"Obsidian is the IDE; the LLM is the programmer; the wiki is the codebase."

Use this template

This repo is a GitHub template. Click Use this template → create your repo → then:

Or install the toolkit only and scaffold locally:

Why not RAG?

Classic RAG retrieves fragments at query time. Nothing accumulates. Ask a question that synthesizes five documents and the LLM rediscovers the pieces every time.

LLM Wiki is different. When you add a source, the agent reads it, extracts key information, and integrates it into a persistent wiki — updating entity pages, revising synthesis, flagging contradictions. Knowledge is compiled once and kept current.

See it in action

Terminal demo — CLI search, ingest status, lint, and expand in ~30 seconds:

Play locally: asciinema play assets/demo.cast (after pip install asciinema)

Ingest flow — one source update touches the whole wiki:

Demo graph — two sources link to 15+ pages (examples/demo/):

From a repo clone

Note: This repository is the toolkit, not a wiki root. Example wikis live under examples/. Always pass --root or use make shortcuts:

Open examples/demo/ as an Obsidian vault to explore the live graph.