AI native but doesn't feel production ready, I've kept encountering random bugs. I still use it over n8n for prototyping LLM pipelines quick