SurrealDB

What SurrealDB does

SurrealDB is an AI-native, multi-model database built in Rust that unifies various data models (document, graph, relational, time-series, geospatial, key-value) into a single engine. It offers powerful search and retrieval capabilities (vectors, full-text, hybrid) and built-in real-time and event-driven features. The platform aims to simplify tech stacks, reduce complexity and operational overhead, and accelerate application development. SurrealDB serves as a single data and logic layer for AI agents, knowledge graphs, real-time applications (like recommendation engines and fraud detection systems), and OLTP applications requiring multiple data types. It can also function as a backend-as-a-service with direct user authentication and can run embedded due to its single Rust binary nature. The company also offers Spectron, a memory and knowledge layer for AI agents that runs on SurrealDB, providing persistent memory and knowledge graphs.

Products

• SurrealDBAn AI-native, multi-model database built in Rust designed to unify multiple data models into a single, powerful engine. It combines document, graph, relational, time-series, geospatial and key-value data types natively, with powerful search and retrieval (vectors, full-text, hybrid) and built-in real-time and event-driven capabilities.
• SpectronThe memory and knowledge layer for AI agents, providing persistent memory and knowledge graphs. It runs on SurrealDB Cloud and offers a unified, ACID-transactional memory model for agents, handling documents, conversations, entities, attributes, relations, embeddings, and traces.

Key capabilities

• AI-native database
• Multi-model support (document, graph, relational, time-series, geospatial, key-value)
• Built in Rust
• Unified query language (SurrealQL)
• ACID compliance
• Vector search
• Full-text search
• Hybrid search (combining vector, graph, structured filters)
• Real-time subscriptions and live queries
• Event triggers and asynchronous events
• WebSocket streaming
• Distributed storage architecture
• Compute-storage separation
• Horizontal scalability
• Automatic data sharding
• Multi-region replication
• Multi-tenancy data separation (namespaces and databases)
• Schemafull or schemaless data models
• Versioned temporal tables (experimental)
• Table fields with data type enforcement and default values
• Table events for custom logic
• Built-in authentication and access control (RBAC, record-level permissions, JWT auth)
• Deployment options: in-memory, embedded, single node, distributed, cloud (SurrealDB Cloud), self-hosted, Docker, Kubernetes
• Model Context Protocol (MCP) integration
• Memory-safe, thread-safe, and crash-resilient core
• Quorum consensus for distributed storage
• Object storage backed data (S3, GCS, Azure Blob)
• Automated daily and custom managed backups
• Cloud management dashboard
• Team collaboration features
• Workload isolation
• Instance metrics, query logs and traces
• SDKs for Java, JavaScript, Kotlin, Mojo, .NET, PHP, Python, Rust, Swift
• Surrealist UI (visual IDE)
• CLI tools
• REST API
• WASM plugins
• Graph-native query engine
• Rich edges (relationships as documents)
• Co-located embeddings
• Autonomous memory understanding (connection discovery, knowledge consolidation, implicit inference)

Use cases

• AI Agents (generative AI apps, multi-step agents)
• Digital Twins (modeling assets, dependencies, telemetry)
• Knowledge Graphs (modeling relationships, contextual insights)
• Embedded and Edge computing (low-latency, offline-ready workloads)
• Real-Time Applications (recommendation engines, fraud detection, collaborative apps)
• Access control & entitlements (graph-based permissions)
• Infrastructure monitoring
• Customer service AI
• Personalized recommendations
• Influencer marketing
• Loyalty platforms
• Cybersecurity data storage
• Sports analytics
• AI Content Copilots
• Podcasting platforms
• Customer support systems

AI approach

SurrealDB is an AI-native, multi-model database designed to unify various data models (document, graph, relational, time-series, geospatial, key-value) into a single engine. It provides a "context layer for AI agents" by consolidating knowledge, memory, and context. It supports powerful search and retrieval (vectors, full-text, hybrid) and real-time capabilities. The platform, including Spectron, is built to provide persistent memory and structured context for AI agents, enabling them to reason over complete and consistent data, and to coordinate in multi-agent systems. It integrates with various AI frameworks and embedding providers.

Tech named: vector database, Graph RAG, RAG, Vector Database, Multi-Model Database, knowledge graphs, embeddings, full-text search, AI agents, Spectron, SurrealQL, Model Context Protocol (MCP), Rust

Industries served

• Software Development
• Artificial intelligence
• Defence and aerospace
• Gaming and entertainment
• Energy and manufacturing
• Finance and FinTech
• Healthcare
• Retail and e-commerce
• Technology

What it says sets it apart

• Unifies six databases into one multi-model engine (document, graph, relational, time-series, geospatial, key-value) with a single query language and ACID transactions.
• AI-native design, built specifically to consolidate knowledge, memory, and context for reliable production-scale AI agents.
• Eliminates the 'frankenstack' problem by replacing multiple disparate systems (vector DB, graph DB, document store, cache, message broker, memory middleware) with a single, unified platform.
• Spectron provides a unique, ACID-transactional memory layer for AI agents, addressing common failure modes like entity collisions, silent overwrites, scope leakage, and untraceable answers.
• Offers flexible deployment options (in-memory, embedded, edge, self-hosted, fully managed cloud) with the same API across all environments.
• Built in Rust for high performance, reliability, memory safety, and crash resilience.
• Next-generation distributed storage architecture with compute-storage separation and quorum consensus on commodity object storage, avoiding proprietary cloud database storage tiers.
• Real-time capabilities (live queries, events, WebSocket subscriptions) are built directly into the database, eliminating the need for external message brokers or polling.
• Comprehensive security features including RBAC, record-level permissions, JWT auth, multi-tenant isolation, and compliance certifications (SOC 2 Type 2, GDPR, Cyber Essentials Plus, ISO 27001).
• Simplifies context engineering by allowing graph traversal, vector search, structured filters, and temporal queries in a single SurrealQL statement.

Funding rounds we track

From the AI funding tracker — rounds as reported by the linked publications.