rohitg00

LLM Wiki v2

A pattern for building personal knowledge bases using LLMs. Extended with lessons from building agentmemory, a persistent memory engine for AI coding agents.

This builds on Andrej Karpathy's original LLM Wiki idea file. Everything in the original still applies. This document adds what we learned running the pattern in production: what breaks at scale, what's missing, and what separates a wiki that stays useful from one that rots.

What the original gets right

The core insight is correct: stop re-deriving, start compiling. RAG retrieves and forgets. A wiki accumulates and compounds. The three-layer architecture (raw sources, wiki, schema) works. The operations (ingest, query, lint) cover the basics. If you haven't read the original, start there.

retlehs

#!/usr/bin/env bash
set -euo pipefail

DOMAIN="${1:-example.com}"
RELEASE="${CC_RELEASE:-cc-main-2026-jan-feb-mar}"
CACHE="${HOME}/.cache/cc-backlinks/${RELEASE}"
BASE="https://data.commoncrawl.org/projects/hyperlinkgraph/${RELEASE}/domain"

VERTICES="${CACHE}/domain-vertices.txt.gz"
EDGES="${CACHE}/domain-edges.txt.gz"

karpathy

LLM Wiki

A pattern for building personal knowledge bases using LLMs.

This is an idea file, it is designed to be copy pasted to your own LLM Agent (e.g. OpenAI Codex, Claude Code, OpenCode / Pi, or etc.). Its goal is to communicate the high level idea, but your agent will build out the specifics in collaboration with you.

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.

k3zi

Table of Contents

Accent Types　アクセントの種類｜Numbers 数字｜Counters　数え方｜Suffixes｜Verbs 動詞｜Adjectives 形容詞｜Prefixes｜〜もの｜Special Common Exceptions｜Regions, Cities & Places｜Names｜Trends

---

BeadtMODE

{
  "contracts@0.4.0/MyToken.sol": ".deps/npm/contracts@0.4.0/MyToken.sol",
  "@openzeppelin/contracts@5.6.0/token/ERC20/ERC20.sol": ".deps/npm/@openzeppelin/contracts@5.6.0/token/ERC20/ERC20.sol",
  "@openzeppelin/contracts@5.6.0/token/ERC20/IERC20.sol": ".deps/npm/@openzeppelin/contracts@5.6.0/token/ERC20/IERC20.sol",
  "@openzeppelin/contracts@5.6.0/token/ERC20/extensions/IERC20Metadata.sol": ".deps/npm/@openzeppelin/contracts@5.6.0/token/ERC20/extensions/IERC20Metadata.sol",
  "@openzeppelin/contracts@5.6.0/utils/Context.sol": ".deps/npm/@openzeppelin/contracts@5.6.0/utils/Context.sol",
  "@openzeppelin/contracts@5.6.0/interfaces/draft-IERC6093.sol": ".deps/npm/@openzeppelin/contracts@5.6.0/interfaces/draft-IERC6093.sol",
  "@openzeppelin/contracts@5.6.0/access/Ownable.sol": ".deps/npm/@openzeppelin/contracts@5.6.0/access/Ownable.sol",
  "contracts@0.4.0/mbt.sol": ".deps/npm/contracts@0.4.0/mbt.sol",
  "contracts@0.4.0/MyConrract.sol": ".deps/npm/contracts@0.4.0/MyConrract.sol",

BeadtMODE

{
  "contracts@0.4.0/MyToken.sol": ".deps/npm/contracts@0.4.0/MyToken.sol",
  "@openzeppelin/contracts@5.6.0/token/ERC20/ERC20.sol": ".deps/npm/@openzeppelin/contracts@5.6.0/token/ERC20/ERC20.sol",
  "@openzeppelin/contracts@5.6.0/token/ERC20/IERC20.sol": ".deps/npm/@openzeppelin/contracts@5.6.0/token/ERC20/IERC20.sol",
  "@openzeppelin/contracts@5.6.0/token/ERC20/extensions/IERC20Metadata.sol": ".deps/npm/@openzeppelin/contracts@5.6.0/token/ERC20/extensions/IERC20Metadata.sol",
  "@openzeppelin/contracts@5.6.0/utils/Context.sol": ".deps/npm/@openzeppelin/contracts@5.6.0/utils/Context.sol",
  "@openzeppelin/contracts@5.6.0/interfaces/draft-IERC6093.sol": ".deps/npm/@openzeppelin/contracts@5.6.0/interfaces/draft-IERC6093.sol",
  "@openzeppelin/contracts@5.6.0/access/Ownable.sol": ".deps/npm/@openzeppelin/contracts@5.6.0/access/Ownable.sol",
  "contracts@0.4.0/mbt.sol": ".deps/npm/contracts@0.4.0/mbt.sol",
  "contracts@0.4.0/MyConrract.sol": ".deps/npm/contracts@0.4.0/MyConrract.sol",

dabit3

See more of my writing here. Also check out Devin

In this post, I'll start from scratch and build up to OpenClaw's architecture step by step, showing how you could have invented it yourself from first principles, using nothing but a messaging API, an LLM, and the desire to make AI actually useful outside the chat window.

End goal: understand how persistent AI assistants work, so you can build your own (or become an OpenClaw power user).

First, let's establish the problem

When you use ChatGPT or Claude in a browser, there are several limitations:

bradtraversy

Common Terminal Commands

Key Commands & Navigation

Before we look at some common commands, I just want to note a few keyboard commands that are very helpful:

• Up Arrow: Will show your last command
• Down Arrow: Will show your next command
• Tab: Will auto-complete your command
• Ctrl + L: Will clear the screen

davidondrej

AGENTS.md

Project Overview

• Project: [Name] — [one sentence describing what it does]
• Target user: [e.g. developers, small businesses, lawyers]
• My skill level: [beginner / intermediate / expert]
• Stack: [e.g. Next.js, Supabase, Tailwind — or "see package files"]

Commands

• Install: [e.g. npm install, pip install -r requirements.txt]