SurrealDB
Rust-built multi-model database unifying documents, graphs, vectors, and more for AI apps.
About
SurrealDB is a London-based, cloud-native multi-model database built in Rust, designed for real-time and AI-native applications. One engine covers relational, document, graph, time-series, vector, search, geospatial, and key-value data, plus embedded business logic, so teams stop stitching together three or four databases through middleware. Version 3.0 reached general availability in early 2026 with an explicit pitch at AI agent memory and context graphs.
The company has raised about $44 million total, including a $23 million Series A extension in February 2026 that added Chalfen Ventures and Begin Capital alongside FirstMark and Georgian, and it positions SurrealDB as the persistent memory layer for agentic systems.
Honest limitations: multi-model breadth is also a risk; any single model (say, vector search at scale) may trail a purpose-built specialist. The 3.0 rewrite means older 1.x/2.x guides and client libraries can be stale, and as a venture-backed startup its long-term roadmap will follow enterprise demand, not necessarily the open-source community.
What it can do
- One engine for relational, document, graph, vector, time-series, and search data
- Embedded business logic and access control inside the database layer
- Agent memory and context-graph primitives aimed at agentic AI apps
Setup
Run the single binary or Docker image, or sign up for Surreal Cloud; query with SurrealQL.
Prerequisites
None for local use; a cloud account for the managed service.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 81; Trending score 82/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.