Use CrabGraph locally while you build, then move to managed graph capacity when teams, agents, and applications need the same connected context.
Share materialized graph capacity across RAG pipelines, agents, analytics, and application APIs.
For local development, tests, prototypes, and single-service graph workloads.
For teams running shared graph workloads across products, agents, and analytics systems.
For regulated teams that need dedicated support, deployment control, and security review.
CrabGraph is designed around one graph layer serving many consumers. That keeps cost tied to real graph usage instead of every team rebuilding the same relationships.
The API stays stable. The operational surface moves from an embedded dependency to managed infrastructure.
Load a schema file, start CrabGraph from your application, and use a clean local graph in tests and early features.
Move hot traversals to a managed cluster so apps, agents, and analysts can reuse the same graph context.
Add private networking, audit trails, role-based access, and capacity planning as the graph becomes production infrastructure.