Keep your knowledge graph current without ever querying production directly.

PolySQL connects your production databases, the system of record for most of what your company actually does. Strand reads from a replica rather than your primary database, so the graph stays current without adding load to production.

How it works

Provide a read replica connection string and select the tables and schemas to include. Cortex uses change data capture to stream updates as they happen, rather than polling on an interval, so the graph reflects production state with minimal lag. Foreign key relationships in your schema are detected automatically and become edges in the graph.

Configure

The PolySQL connection is configured from the integrations panel, where connection details and table scope are managed per workspace. IP allowlisting may be required depending on your database provider’s network configuration. Schema changes on the source database are detected automatically, and an admin is notified if a change affects existing graph mappings.