How I Built a Local AI Workflow Automation Stack From Scratch

n8n, Claude Code, and a single terminal command — no subscriptions, no vendor lock-in, no UI clicking.

Most iPaaS tools charge per task, lock your workflows inside their platform, and make you click through a UI every time you want to change something. That’s fine until you’re building something that needs to move fast or scale without the bill growing with it.

This setup is different. n8n runs locally on Docker — your workflows, your data, your infrastructure, no usage fees. Claude Code sits inside VS Code and lets you prompt your way through building, debugging, and modifying workflows without context switching between tools.

What used to take an afternoon of visual builder clicking can be done in a conversation.

The result is a stack that’s faster to build on, cheaper to run, and actually yours to control.

The full setup guide is below. It’s written so anyone can follow it from a blank machine.