Vector search answers “what is similar?”, but agents need “what is related?”
NeuroIndex combines vector search, semantic graph recall, and persistent memory in a fully local, open-source project.
Would love feedback on design, use-cases, and improvement ideas.