The idea is:
Hi everyone! I have an idea (and something that would be super useful for everyone—including myself): to create a “control” node (let’s say an AI-powered Switch node that, based on the user’s text, knows exactly where to send the request) to “route” to the right AI Agent.
The idea is to have specialized agents for different tasks, each “connected” to certain tools that the others aren’t (and each agent can have its own specific AI—like one using GPT-4.1 for complex things, another with Gemini Flash for others, etc.—so we can save money, tokens, and avoid using an expensive AI for trivial tasks). They would all share the “same” conversation memory (so they know where the user is in the process) and respond based on that.
example:
My use case:
To be able to use the right AI models at the right moment: advanced reasoning when it’s needed, simple answers when that’s enough, etc.
I think it would be beneficial to add this because:
FlowiseAI already has this, and it works.
Any resources to support this?
Not for now
Are you willing to work on this?
Yes