Matthew Harris Byteflux

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.

	<?php

	// Your GitHub username goes here.
	$user = $_ENV['GITHUB_USER'] ?? '';

	// If using 2FA, you must create a Personal Access Token at https://github.com/settings/tokens
	// The access token can be used in place of the password and should be revoked when no longer needed.
	$password = $_ENV['GITHUB_PASSWORD'] ?? '';

	// This is the base URL to the GitHub API.

	import sys
	import os
	import math
	import urllib2
	import json
	import time
	import shutil
	import uuid
	from nbt import nbt # pip install nbt