Browse / LLM wiki / Second brain / llm-wiki-template
llm-wiki-template
max-clearframe
A pattern for building personal knowledge bases using LLMs
Open max-clearframe/llm-wiki-template → Checked against the four rules.
Works with any agent
ships AGENTS.md, the cross-tool instruction file.
In its own words
A pattern for building personal knowledge bases using LLMs.
Credit: This idea originates with Andrej Karpathy, who published it as a GitHub gist in April 2026. This repo is a ready-to-use template built from that concept. See the original gist.
Clone this repo, open it in your LLM agent, and start ingesting sources. The agent writes and maintains the wiki. You do the reading, the sourcing, and the thinking.
Quick Start
- Clone this repo
- Open the directory in your LLM agent (Claude Code, OpenAI Codex, or any agent that reads
CLAUDE.md/AGENTS.md) - Drop a source file into the appropriate
raw/subfolder - Tell your agent:
ingest [filename] - Browse the results in
wiki/
The agent reads the schema in CLAUDE.md (or AGENTS.md) and knows exactly what to do.
The Core Idea
Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up. NotebookLM, ChatGPT file uploads, and most RAG systems work this way.
Quoting 14% of the README. Read the rest on GitHub ↗
Read from the repository's git tree on 2026-08-17.