Let's say I have a workflow and I want it to split and go 3 different ways AT THE SAME TIME. Imagine an LLM receives a request and produces 3 json objects that need to go 3 different paths and can be processed concurrently. How would you do that and, as a follow up, is there a way to join those 3 paths back together at the end?

Has anyone found a way to use the new MCP server connection to debug localhosted workflows with the cloud version of Claude or Chat GPT? How did you manage to solve it?

I just got this email and was curious to what it was. From a quick glance seems like digital ocean is taking out the hassle of self hosting and basically creating premade packages for n8n hosting. Is that right? What are your thoughts?