Harper (formerly HarperDB)
activeby HarperDB, Inc. · since 2017 · also known as HarperDB, Harper, Harper Pro, Harper Fabric
Node.js runtime that fuses database, cache, app logic and MQTT/WebSocket messaging in one process.
Overview
Harper, called HarperDB until a 2025 rebrand, is a Node.js platform that runs the database, cache, application code and messaging in a single process so requests never cross a network hop. Tables are declared in GraphQL schema files and exported as REST endpoints automatically, data lives in LMDB or, since 5.0, RocksDB, and it adds secondary and HNSW vector indexes, caching with expiry, blob storage and publish/subscribe over MQTT, WebSockets and server-sent events. Harper 5.0 (April 2026) moved the core to Apache 2.0, while Harper Pro adds multi-node replication and certificate management under the Elastic License 2.0; Harper Fabric is the managed, geo-distributed offering. It fits latency-sensitive edge APIs and caches better than general relational workloads.
Strengths & trade-offs
Strengths
- +Database, cache, application code and messaging share one process and memory space
- +Schema annotations generate REST endpoints with no boilerplate
- +Built-in MQTT and WebSocket pub/sub without a separate broker
- +LMDB and RocksDB storage engines with HNSW vector indexes
- +Apache 2.0 core since Harper 5.0
Trade-offs
- −Replication and clustering sit in the ELv2-licensed Harper Pro, not the open-source core
- −Application logic must be written in JavaScript/Node.js
- −SQL support is limited compared with relational databases
- −Smaller community than Redis, MongoDB or Postgres
Great for
- +Edge caching and API acceleration in front of slower origins
- +Real-time apps needing MQTT or WebSocket fan-out
- +Geo-distributed application backends
- +Vector lookups next to application logic
Think twice for
- −Complex relational reporting
- −Teams standardised on non-JavaScript backends
- −Large analytical scans
Capabilities
Native or official first-party support · 10 of 25 tracked
Scale & consistency
- Comfortable data size
- Medium · ≤ 10 TB
- Horizontal scaling
- Addon
- Multi-region
- Active active
- Transactions
- Limited ACID
- Consistency
- Eventual, Strong
- Schema
- Flexible
- Architecture
- Client server, Distributed, In memory
- Query languages
- REST query parameters, GraphQL, SQL (limited)
Single-node core; multi-node and geo-distributed replication come with Harper Pro or Harper Fabric.
Deployment
Ecosystem
Alternatives to Harper (formerly HarperDB)
The in-memory data structure server that became the default cache, now with search, JSON and vectors built in.
harperdb vs redis →Open-source backend in one Go binary: SQLite database, auth, files, realtime and admin UI.
harperdb vs pocketbase →Open-source Postgres backend platform: database, auth, storage, realtime and edge functions in one project.
harperdb vs supabase →Reactive TypeScript backend with a transactional document database, server functions and live queries.
harperdb vs convex →The default document database: flexible BSON documents, rich queries and built-in sharding.
harperdb vs mongodb →Single-threaded stateful actors at the edge, each with its own strongly consistent SQLite store.
harperdb vs cloudflare-durable-objects →Harper (formerly HarperDB) FAQ
What is Harper (formerly HarperDB)?
Harper (formerly HarperDB) is a document database by HarperDB, Inc., first released in 2017. Harper, called HarperDB until a 2025 rebrand, is a Node.js platform that runs the database, cache, application code and messaging in a single process so requests never cross a network hop. Tables are declared in GraphQL schema files and exported as REST endpoints automatically, data lives in LMDB or, since 5.0, RocksDB, and it adds secondary and HNSW vector indexes, caching with expiry, blob storage and publish/subscribe over MQTT, WebSockets and server-sent events. Harper 5.0 (April 2026) moved the core to Apache 2.0, while Harper Pro adds multi-node replication and certificate management under the Elastic License 2.0; Harper Fabric is the managed, geo-distributed offering. It fits latency-sensitive edge APIs and caches better than general relational workloads.
Is Harper (formerly HarperDB) open source?
Yes. Harper (formerly HarperDB) is open source. License: Apache License 2.0 (core); Elastic License 2.0 (Harper Pro).
Does Harper (formerly HarperDB) support ACID transactions?
Partly. Harper (formerly HarperDB) offers limited ACID guarantees (for example, single-record or single-partition transactions), with eventual, strong consistency.
Does Harper (formerly HarperDB) scale horizontally?
Harper (formerly HarperDB) scales horizontally through an official extension, proxy or edition rather than in the core engine. It is comfortable at medium scale (≤ 10 TB). It supports active-active multi-region writes.
How can I run Harper (formerly HarperDB)?
You can self-host it, use the vendor's managed cloud and run it at the edge. Managed offerings include Harper Fabric.
What is Harper (formerly HarperDB) best for?
Harper (formerly HarperDB) is a strong fit for edge caching and API acceleration in front of slower origins; real-time apps needing MQTT or WebSocket fan-out; and geo-distributed application backends. Think twice for complex relational reporting.
What are the best alternatives to Harper (formerly HarperDB)?
Common alternatives to Harper (formerly HarperDB) include Redis, PocketBase, Supabase, Convex and MongoDB. The right choice depends on your workload, scale and hosting; WhichDB's smart match ranks them against your requirements.
How popular is Harper (formerly HarperDB)?
Harper (formerly HarperDB) ranks #317 of 844 databases on the WhichDB Score (9.1/100), and #30 among document databases. The score combines public adoption signals such as package downloads, GitHub activity, job posts and developer discussion.