NodeDB

NodeDB is a universal database engine designed to consolidate multiple data models and workloads into a single system. It addresses the operational complexity faced by AI product teams and developers who typically deploy and integrate several specialized databases—such as relational, vector, graph, key-value, full-text search, and scientific array stores—by offering all these capabilities natively within one binary. NodeDB aims to eliminate the need for glue code, multiple deployments, and cross-service synchronization by fusing these engines at the database layer.

The platform supports relational tables, vector indexes with HNSW and PQ, property graphs with thirteen algorithms, key-value storage with constant-time hash lookups, full-text search using BM25, and ND sparse arrays for scientific data such as genomics and climate observations. It also provides features like bitemporal storage for audit, time-travel, and GDPR-compliant erasure, as well as row-level security, role-based access control, audit logging, and tenant isolation for multi-tenant SaaS scenarios. NodeDB integrates these engines with a unified SQL planner, enabling cross-engine queries in standard SQL. Example queries include combining vector search with metadata pre-filtering, hybrid BM25 and vector ranking, and spatial plus vector operations within a single statement.

NodeDB is delivered as a single Rust binary and is compatible with existing PostgreSQL clients, allowing users to connect without changing their tooling. The database is positioned in public preview and is already in use in real projects. NodeDB is intended for teams building AI products or applications that require diverse data models and seek to simplify deployment and reduce operational overhead by unifying multiple database functionalities.