Semantic Models (Beta): Author a single, governed analytical layer on top of your physical schema and forward-engineer it into Snowflake Semantic Views or Cort…Semantic Models (Beta): Author a single, governed analytical layer on top of your physical schema and forward-engineer it into Snowflake Semantic Views or Cortex Analyst–compatible YAML. Define metrics, dimensions, facts, and filters once so every dashboard and AI engine uses one consistent definition. Works across all database types; never modifies the physical model.
- Two-layer design — set project-wide semantic defaults once; named models inherit them and override only the deltas (Reset to default / Push changes to default)
- Classify fields as Dimension, Time Dimension, or Fact; configure synonyms, SQL expressions, sample values, and access modifiers
- Model- and table-level metrics, table-scoped filters, verified NL-to-SQL queries, and custom AI instructions
- Semantic joins (Left/Right/Inner/Full/AsOf) from inherited or local relationships
- Forward Engineer to DDL (CALL SYSTEM$CREATE_SEMANTIC_VIEW_FROM_YAML) or Semantic YAML
Beta: no Reverse Engineering, branch merge, logical-to-physical, or Databricks metrics views yet.