It is "experimental". The AI sometimes produces code that doesn't fully work (this is part of the fun). Refresh and try again. It works 90% of the time I would say.
Here I've created a basic Paint program with no pens, brushes or other drawing tools. Instead you get the AI to create the painting tools for you and it writes, tests and deploys the JavaScript for you live, usually within 30 seconds.
It only uses OpenAI's gpt-3.5-turbo model too which is fast and good enough for this use case.
We’ve used N8N for about 2 years, at scale, in production, self-hosted on a VM with docker compose. It’s phenomenal.
We run every piece of client-specific custom integration work through it.
Plumbing a few systems together is a piece of cake - either triggered by webhook or scheduled cron-like.
There are tonnes of out the box nodes for common services, but for everything else the JavaScript code block and HTTP request nodes fill in any missing gaps.
We’ve tried Zapier, Integromat and loads of others - N8N has nailed the UX for us.