Beyond the Chatbox

A field guide for people who want to stop chatting with AI and start working with it — in their codebase, their knowledge base, and their workflows.

Written for technically curious folks who currently live in Claude.ai, ChatGPT, or Gemini and want to understand what's possible when you bring AI into your actual work environment.

Beyond the Chatbox — moving from web chat to working with AI in your files, knowledge, and workflows

AI agents and assistants: start with CLAUDE.md and AI-SETUP-PROMPT.md.

Who this is for

Techs and IT professionals expanding their AI skill set
People who've used web-based AI chat tools and want to go deeper
Anyone curious about context engineering, IDE-based AI, or automation workflows
No prior coding experience required to get started

Prerequisites

Start here first: 00 — Prerequisites — Claude Pro, Git, VS Code, and Docker Desktop. Takes 30-60 minutes. Do this before anything else.

Sections 01-03: Just a browser. Read on GitHub directly.
Sections 04+: Prerequisites must be installed.

Learning Path

Work through these in order. Each section builds on the last.

Section	Topic	What you'll learn
00 - Prerequisites	Setup	Claude Pro, Git, VS Code, Docker Desktop
01 - Foundations	AI fundamentals	How LLMs work, key concepts, best resources
02 - GitHub Basics	Git and GitHub	Version control, repos, PRs, how AI teams use GitHub
03 - AI Tools Landscape	Tools overview	Web chat vs IDE-based AI — when each makes sense
04 - IDE Setup	Getting started	Install Claude Code, set up VS Code, scaffold your first project
05 - Context Engineering	The core skill	CLAUDE.md, repo structure, prompting patterns
06 - Extended Knowledge	Knowledge system	A markdown-based knowledge folder Claude Code reads as context
07 - Repo Template	Ready-to-use template	Copy this, fill in your details, have a context-engineered repo in minutes
08 - Automation Platforms	No-code workflows	n8n, Make, Zapier — build real automations without writing code
09 - MCP Servers	Extend Claude's tools	Build a working Python MCP server that gives Claude new capabilities

Quick Start

New to this and not a coder? Open AI-SETUP-PROMPT.md. It walks you through getting this guide in front of an AI assistant (a browser Claude window works) and pasting one prompt that gives you a personalized starting point. No tools required to begin.

How to use this repo

Option A — Read it on GitHub. No setup required. Start at 01-foundations/README.md and work through the sections in order.

Option B — Claude Desktop (no terminal). Install Claude Desktop, set up the Filesystem and GitHub connectors, then paste the bootstrap prompt into a conversation to scaffold your first project. No Node.js, no command line.

Option C — Claude Code CLI (full power). Complete section 04's CLI path (Node.js + npm install -g @anthropic-ai/claude-code), then open this repo in Claude Code and run the bootstrap prompt. Better for developers who want Claude to run commands and manage git autonomously.

The core idea

Most people use AI as a search engine with a chat interface. Context engineering is the practice of structuring your work so the AI always has the right information at hand — without you having to re-explain everything every session.

This repo teaches you how to do that.

Contributing

This is a living document. If you find outdated links, have a better resource to suggest, or want to add an example workflow, open a pull request. See CONTRIBUTING.md for the house style and how to submit.

License

MIT — free to use, copy, adapt, and share.

beyond-the-chatbox