The idea is:
Apologies if there is already a request for this, but I did not see one after doing a quick search in the ‘feature requests’ category.
The title says it all – would love to have the n8n AI Assistant (or similar functionality) available to local n8n installs. It appears that this is the direction that the industry is heading, as other automation platforms are beginning to integrate ‘AI assistant’ functionality for creating workflows. Overall though I like n8n much better than all the rest!
I realize that the cloud version of n8n already has this something like this built in. But I prefer to self host n8n, for various other reasons. That said, I’d be more than happy to pay for this functionality IF it was ever available for local n8n installs!
I recall reading somewhere a while back (either in the main discussion, or maybe on Reddit?) that the reason the n8n devs have not done this yet is because they want to be able to control the logic and code for the ‘AI assistant’ functionality, server-side (and perhaps implement a custom-trained LLM or RAG, etc). That’s fair, but I would imagine that are still ways that this could be implemented for local installs. Perhaps via a dedicated API? Or maybe a browser plugin?
I think that some of the better existing generative AI platforms (ChatGPT, Claude, Gemini, etc) will be able to spit out perfect n8n workflows from a text prompt fairly soon, but after spending hours recently testing this personally, they’re definitely not quite “there” yet!
My use case:
Type a text prompt and then n8n automatically creates a fully functional workflow. Of course, the more complex a workflow is, the more iterations would likely be needed by the AI assistant before it “gets it right”. But it would still be much faster than building workflows manually.
I think it would be beneficial to add this because:
This would obviously save a TON of time when creating workflows, and would make n8n more accessible to waaaaay more less “techy” users (= more $$ for n8n).
Any resources to support this?
Are you willing to work on this?
I’m not much of a coder, but I’d be happy to contribute in whatever other way I could (UI design, logic, documentation, etc).