Great question. It’s hard to boil down that way, I think, because the answer depends on the user’s perspective and needs.
However, from an n8n perspective, focusing solely on n8n might mean missing out on a couple of specialized areas that Flowise brings to the table for no-code, flow-based systems.
As you know, n8n is very flexible and offers a wide array of pre-built integrations. Flowise is more specialised in building and handling AI-centric workflows and AI Agents.
For example, Flowise offers a more sophisticated implementation for vector store record management, which I find critical for building nocode/low-code solutions.
Handling and upserting new documents to a vector store, either based on uploaded documents or web scraped content, is way easier in Flowise compared to n8n.
The UI endpoint - a chatbot - as provided by Flowise, is more configurable and ‘production ready’ than the one we get from n8n’s Chat Trigger node. I think it still needs some iterations and feature enhancements to really do the job.
When you build a chat flow in Flowise, an API is automatically made available to that flow. This makes it very easy to integrate. However, n8n has way more nodes/integrations than Flowise. But how many of these do you actually use?
Flowise is optimized for AI chatflows and AI Agents so every step in making this work is designed with this purpose (record management, memory, vector store integrations etc.) It’s not “build on top” as it is in n8n.
The number of releases and speed of development of n8n is way better than that of Flowise. The user community is bigger and it’s easier to get help to solve issues or get inspiration on flow design.
If a node is being updated in Flowise, it’s extremely easy to update an existing chatflow after deploying a new release. Just a click of a button in a flow, and all updated nodes in that flow are automatically being updated to the latest release. This feature I really miss in n8n!
In short, if you only studied n8n, you might be overlooking a platform that is better optimized for advanced AI data handling, which could be a significant advantage depending on your project’s requirements. Though, n8n is developing in this direction, I think, but I still miss simple features like record management, improved chatbot features, ease of handling document/data upsertion to vector stores, ease of handling uploaded documents to a chat (this is unnecessarily complex today at n8n!)
I find it a bit hard to get new AI feature requests prioritized in n8n development pipeline, as these features compete against other more general automation features. I hope n8n will improve their AI focus even more in the near future to address obvious needs like what I describe above.
You do such a terrific job at n8n!