ARCHITECTURE
UK multi-modal and multi-model database startup SurrealDB has announced v4.0 with a new storage engine, a SurrealDB Earth offering running it serverless across Cloudflare, and a distributed SurrealDS running it on node clusters backed by object storage.
The company says that it’s a database and memory for every AI agent, supporting vector, graph and document data. It’s made no formal news announcement, relying instead on webpages, a blog post, and X social media posts.
Co-founder and CEO Tobie Morgan Hancock blogs: "The goal is one database that grows from a single lightweight workload to the infrastructure behind millions of users, devices and agents. SurrealDB 4 is the biggest step towards it so far."
SurrealDB says:
- SurrealDB 4 is the database engine: a rebuilt query core, PostgreSQL SQL and MySQL support over their wire protocols, mql.js 1.0 and more than 240 fixes that no 3.x release contains. It runs embedded, on a laptop, in the browser, on your own servers or in SurrealDB Cloud.
- SurrealDB Earth is serverless SurrealDB, built on Cloudflare, that scales to as many agents and users as you need, wherever they are. Every Earth database is its own isolated instance, with dedicated storage and commit log.
- SurrealDS is distributed storage backed by object-storage for large SurrealDB clusters, in SurrealDB Cloud multi-node clusters and self-hosted Enterprise Edition deployments.
- Agent Memory, launched in October, is a separate, managed service built on the same engine. Memory in one query is built into the engine, and Agent Memory is the managed way to get it.
The v4 database engine is a much better performer than the v3 version with indexed queries projected to be ~22× faster, full-table scans ~2.3× faster, CRUD (Create, Read, Update, and Delete operations)~42% faster (crud-bench). Vector indexes build much faster (e.g., 1 million × 768-dimension HNSW 7.4x faster (986s s ~7,290s), and filtered vector search drops from ~27 ms to ~1.1 ms, with SIMD scoring and other optimizations. Full-text ranking is also improved.
There is a unified query executor which replaces multiple previous planners. The engine is compatible with PostgreSQL and MySQL wire protocols.
SurrealDB Earth runs on Cloudflare’s global points of presence. It has isolated databases per application or agent, spun up and down on demand, from zero to 100,000 transactions per second back to zero, for example. Users pay for requests and storage, with no idle server costs. Data is durable in Cloudflare storage.
SurrealDS is distributed storage for workloads that grow beyond a single machine and runs on object-backed storage. It separates compute from durable storage and each can scale separately. Query nodes keep the state and caches they need to run queries quickly, while durable database pages live in shared object storage such as Amazon S3, Azure Blob or GCS.
V3.x users can upgrade but SurrealDB recommends they wait until v4.1 which will be more robust and have upgrade bugs ironed out. It notes the v3.x to v4 migration is one-way, meaning migrated datastores cannot be reopened with v3.x.
For detailed instructions, including how to enable the experimental upgrade, compatibility considerations and rollback options, see its 3.x to 4.0 upgrade guide.
Bootnote
Checkout SurrealDB benchmark performance here. Note the engine is free to self-host under BSL 1.1; the SDKs and tools are open source under Apache 2.0 or MIT.