1 connected nodes across dictionary, workflows, comparisons, prompts, tool stacks and use cases.
A cache that returns a stored LLM answer when a new question means the same thing as an earlier one, even if worded differently.