SurrealDB

A self-hosted multi-model database for AI applications that combines graph, vector and full-text search and can run embedded or in a browser.

SurrealDB is a database for developers building AI agents, knowledge graphs and applications with several kinds of data. It lets application records and retrieval data share one engine.

Multiple data models: Store documents, graphs, vectors, relational records and time-series data, and query them together within an ACID transaction.
Combined retrieval: Use SurrealQL to bring graph traversal, vector similarity, full-text search and filtering into one query for applications such as GraphRAG.
Access controls: Built-in authentication and role, record and field permissions govern what users, services and agents can read or change.
Live updates: Live queries and change streams deliver data changes to connected applications as writes commit.
Client and AI access: Query through REST or WebSockets, or connect models that support MCP to the database.

Built in Rust, SurrealDB can run on your own server, inside an application as an embedded database, or in a browser through WebAssembly. It supports single-node deployments and distributed clusters. SurrealDB Cloud is a separate managed hosting service with a choice of cloud provider and region.