A local-first persistent memory server that uses Voronoi partitioning and temporal intelligence to provide LLMs with long-term context across different sessions and clients. It features automatic drift detection to prevent hallucinations by identifying and penalizing outdated information as project
View & run on Rokha →