Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
minDBgithub.comminDB is an extremely memory-efficient vector database that uses a highly compressed search index combined with disk-based reranking to achieve high recall and low latency while using significantly less RAM than traditional HNSW-based systems. It can index and query 100 million 768-dimensional vectors with peak memory usage of around 3GB, enabling local processing of large datasets like Wikipedia on an average laptop.
Parse Score
