Create a scriptable “meta-node” that will produce a json node description as output, and that json is the configuration for the node that will be executed. Whether a meta-node can recursively produce meta-nodes ad output might be unnecessary.
While developing a Telegram bot, I noticed that the configuration is a bit too rigid for my purpose: I can add the inline keyboard to a message, but the keyboard definition is rather static: the number of rows and buttons is pre-defined in the parameters, I can indeed use expressions there, but I might need a variable number of lines and buttons as well.
Using a meta-node, I could write a node that takes the output of the previous node and produces an ad hoc configuration that does precisely what I need, in a more flexible way than it is currently possible.
it brings the automation to an higher level, leveraging the existing nodes and increasing the behavior that can be controlled programmatically. On some levels, it might also make the workflows more compact, but could make it more complex. In general, would be a “power user” feature, I think.
I couldn’t find anything like this in the documentation or in the nodes, so I am not sure if this functionality is actually provided in some other way.
Not available to work on this, but I think this might require some core changes since this might introduce some dynamism that n8n is not able to support yet. Not sure about this, just speculating.