Agentic mesh with native graph and vector data
Industry:
Artificial intelligence
Outcome:
Unified agent memory with multi-tenancy and sync
DigitalKin built an agentic mesh ecosystem that unifies document, relational, and real-time state in SurrealDB. The platform now delivers multi-tenant isolation, live updates, and a clear path to graph and vector memory.
Key challenges
Multi-tenant isolation
Customer data boundaries required isolation without spinning up separate clusters.
Real-time state
Polling created laggy dashboards and brittle webhook chains across agents.
Schema agility
Fast-moving workflows needed schema-optional modelling without bottlenecks.
Path to graph data
The team needed an upgrade path to graph relationships without a rewrite.
Solutions
Agentic mesh with graph and vector data
SurrealDB powers search, relationships, and memory with native graph and vector capabilities.
Schema flexibility for turnkey agent data
Provided a database-within-a-database model for agent developers to store data freely.
Live queries for push-based updates
Streamed changes to keep agents and enterprise UIs aligned without polling.
Namespace tenancy and hybrid deployments
Isolated tenants with namespaces and kept parity between cloud and self-hosted nodes.
Results
Simplified multi-tenant stack
SHARED
Namespace isolation reduced per-tenant overhead while preserving boundaries.
Faster iteration loops
↑ VELOCITY
Surrealist accelerated query design, schema evolution, and debugging.
Real-time experience
LIVE
Push-driven UX kept the agent mesh and dashboards in sync.
Graph & vector ready
READY
Prepared to consolidate agent memory and knowledge with relationships.
