Jazz is a local-first relational database and synchronization framework designed for offline-capable applications. It functions as a reactive state management system that treats database records as the primary source of truth, automatically updating user interfaces in real time as underlying data changes.
The project distinguishes itself through a collaborative data synchronization model that utilizes row-level versioning to track branching edit histories. It implements a security engine based on identity-claim row security, which enforces granular permissions on individual records, and supports end-to-end field encryption to protect sensitive data at the source.
The system covers a broad range of distributed data capabilities, including relational schema definition via a domain-specific language, durable stream-based synchronization between local replicas and remote storage, and geo-optimized distributed hosting. It provides a programmable interface for standard database mutations and queries while ensuring consistency across different devices and environments.