SurrealDB

SurrealDB is a multi-model database for AI agents, unifying documents, graphs, vectors, time-series and relational data in one ACID transaction.

AI 개발 도구 AI 지식 그래프 기타

웹사이트 방문

Overview

SurrealDB is a multi-model database positioned as a context layer for AI agents. The product site describes it as one database for documents, graphs, vectors, time-series, relational data, full-text search, and agent memory, all handled within a single ACID transaction.

The platform is designed to reduce the glue code and system sprawl that come from stitching together separate databases, caches, search tools, and middleware. SurrealDB Cloud, self-hosted deployments, and Enterprise options are presented for teams that need prototypes, production workloads, or private-cloud control.

Core capabilities

Native multi-model storage

Documents, graphs, vectors, time-series, and relational data are presented as native primitives in one database, so context does not need to be split across separate systems.

Single transactional snapshot

The site emphasizes one ACID transaction boundary for reading and writing context, which helps avoid partial updates across related data.

Search and retrieval in one engine

Built-in full-text search, vector search, graph traversal, and temporal queries support retrieval workflows for agent memory and hybrid search.

Database-level security controls

Built-in access controls include RBAC, record-level permissions, JWT auth, and multi-tenant isolation, with enterprise compliance options also listed.

Real-time and event-driven features

SurrealDB supports real-time subscriptions, event triggers, and streaming updates for reactive application behavior.

Multiple query and client interfaces

The product pages describe multiple access paths, including SurrealQL, GraphQL, PostgreSQL wire protocol, and MCP server support.

Where it fits

Agent memory and context
Use SurrealDB to keep user history, preferences, embeddings, and relationships in one transactional layer so agent responses can draw from a complete snapshot.
Hybrid search and RAG
Build retrieval flows that combine vector search, full-text search, graph traversal, and temporal filtering without passing data through separate services.
Operational applications with mixed data
Store relational records, documents, and graph edges together when application state spans multiple models and must update consistently.
Real-time product experiences
Use the real-time subscriptions, event triggers, and streaming updates to power reactive products that need immediate changes reflected in the UI or workflow.
Teams with different deployment needs
Choose cloud, self-hosted, or private-cloud deployment when you need a managed start, on-premises control, or an enterprise operating model.

Pros and Cons

Pros

Combines multiple data models in one ACID database instead of splitting context across separate systems.
Supports agent-oriented workflows such as memory, retrieval, graph reasoning, and hybrid search.
Offers cloud, self-hosted, and enterprise deployment paths.
Includes built-in access control and real-time data features.
Provides multiple interfaces and protocols, including SurrealQL, GraphQL, PostgreSQL wire protocol, and MCP server support.

Cons

The public source does not provide a complete integration catalog or connector list.
Some enterprise security and support items are only described at a high level, with specific availability depending on plan or coming soon notes.

FAQ

What is SurrealDB used for?

SurrealDB combines documents, graphs, vectors, time-series, relational data, full-text search, and access control in one database. The source describes it as a single ACID transaction boundary with no middleware for agent memory and context.

Does SurrealDB offer both cloud and self-hosted options?

The site describes SurrealDB Cloud, self-hosted deployments, and Enterprise. Cloud plans include Start and Scale, while Enterprise is for on-premises or private-cloud deployments with custom pricing.

Does SurrealDB support real-time application workflows?

Yes. The product pages highlight built-in real-time subscriptions, event triggers, and streaming updates for reactive experiences.

What support options are available?

The pricing page lists support tiers from standard and business-critical support through premium support with 24x7 coverage. The product pages also mention community support and Surreal Sidekick AI copilot as part of the support experience.

What integrations or interfaces are documented publicly?

The source does not list a full public integration catalog. It does show a MCP server, GraphQL support, PostgreSQL wire protocol support, and client/runtime options across several languages.