Orionbelt Semantic Layer

OrionBelt compiles declarative YAML semantic models into dialect-specific SQL across eight databases including BigQuery, Snowflake, DuckDB, and Postgres. You define dimensions, measures, and business logic once, then query by business concept instead of raw schema. The MCP server exposes compilation and execution tools that let Claude translate natural language into governed SQL without writing dialects by hand. It also runs a pgwire endpoint and Flight SQL server, so you can connect BI tools directly to the semantic layer. The companion OrionBelt Analytics server generates RDF ontologies from database schemas to help AI navigate relationships and avoid fan traps. Best for teams standardizing analytical SQL across platforms or building agentic data access with semantic guardrails baked in.