notcod

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.

Please see the most up-to-date version of this protocol on my blog at https://darencard.net/blog/.

Automatically push an updated file whenever it is changed

Linux

Make sure inotify-tools is installed (https://github.com/rvoicilas/inotify-tools)
Configure git as usual
Clone the git repository of interest from github and, if necessary, add file you want to monitor
Allow username/password to be cached so you aren't asked everytime

Node.js Deployment

Steps to deploy a Node.js app to DigitalOcean using PM2, NGINX as a reverse proxy and an SSL from LetsEncrypt

1. Sign up for Digital Ocean

If you use the referal link below, you get $10 free (1 or 2 months) https://m.do.co/c/5424d440c63a

2. Create a droplet and log in via ssh

I will be using the root user, but would suggest creating a new user

	const puppeteer = require('puppeteer');
	const nodemailer = require('nodemailer');

	class Webpage {
	static async generatePDF(url) {
	const browser = await puppeteer.launch({ headless: true }); // Puppeteer can only generate pdf in headless mode.
	const page = await browser.newPage();
	await page.goto(url, { waitUntil: 'networkidle', networkIdleTimeout: 5000 }); // Adjust network idle as required.
	const pdfConfig = {
	format: 'A4',

	.rounded-corners-gradient-borders {
	width: 300px;
	height: 80px;
	border: double 4px transparent;
	border-radius: 80px;
	background-image: linear-gradient(white, white), radial-gradient(circle at top left, #f00,#3020ff);
	background-origin: border-box;
	background-clip: padding-box, border-box;
	}

	<?php
	<?php
	///amo integra begin

	$values = $hook->getValues();
	$subject = $modx->getOption('emailSubject', $formit->config, 'form-'.$modx->resource->get('emailSubject'));
	$name = $values['name'];
	$phone = $values['phone'];
	$email = $values['email'];

	<?php

	/*
	参考自：
	http://darklaunch.com/2010/09/01/http-status-codes-in-php-http-header-response-code-function
	http://snipplr.com/view/68099/
	*/

	function HTTPStatus($num) {
	$http = array(

	/*
	This smartcontract is used to store documents text on the Ethereum blockchain
	and to get the document by document's hash (sha256).

	*/

	contract ProofOfExistence{

	/* ---- Public variables: */
	string public created;

	import web3

	w3 = web3.Web3(web3.HTTPProvider('https://geth.golem.network:55555'))
	tx = w3.eth.getTransaction('0x00ee93fd1e6fdbdf841b4458a078210449dce89525502c965fd3534ba397e2e3')
	tx.hash

	from eth_account._utils.signing import extract_chain_id, to_standard_v
	s = w3.eth.account._keys.Signature(vrs=(
	to_standard_v(extract_chain_id(tx.v)[1]),
	w3.toInt(tx.r),