Self-hosting n8n used to mean a Docker Compose file, a pile of environment variables, and a sandbox you had to wire up yourself. Now it’s one command.
This video walks through installing n8n locally from scratch, updating an
existing instance, and turning on the AI Assistant with your own model API key.
The command:
curl -fsSL https://get.n8n.io | sh
Update command used:
curl -fsSL https://get.n8n.io | sh -s – --upgrade --version 2.35.3
That single script installs n8n via Docker Compose and pre-configures the code sandbox and the private web search the AI Assistant uses.
Both run on your own machine, and n8n itself is the free Community edition, so the only thing you pay for is the model provider you plug in. Bring an Anthropic, OpenAI, or OpenRouter key and you get the same AI Assistant experience Cloud users have had since July: describe a workflow in plain words and watch it get built, tested, and fixed.
All you need installed first is Docker.