HarperDB

About HarperDB

HarperDB (Harper) is a unified runtime that combines a database, cache, application logic, and messaging into a single process, with the stated goal of eliminating the separate services typically wired together in a backend stack. It provides schema-driven REST and GraphQL APIs generated automatically from table definitions, including filtering, sorting, joins, and pagination, along with built-in support for real-time data through MQTT, WebSocket, and SSE, and geo-distributed replication across regions.

The runtime includes built-in agent memory, embeddings, semantic cache, and context resolution, positioning it for building AI agents that need state and retrieval without assembling a separate vector database, cache, and app server. It also supports vector search and blob storage from the same schema-driven interface used for standard data access.

Deployment options include managed cloud, edge, on-premises, and regulated environments, with the same runtime running consistently across these targets via Harper Fabric. It is intended for developers and coding agents building full-stack applications, APIs, e-commerce systems, and streaming or agentic workloads who want a single system to build, deploy, monitor, and debug rather than coordinating multiple separate services.

Key features

• Unified runtime combining database, cache, application logic, and messaging in one process
• Automatic REST and GraphQL API generation from schema definitions
• Built-in real-time subscriptions via MQTT, WebSocket, and SSE
• Geo-distributed replication with automatic failover across regions
• Built-in agent memory with embeddings, semantic cache, and context resolution
• Deployable to managed cloud, edge, on-prem, or via Docker and Red Hat OpenShift

Frequently asked questions