Connectors

Sema reads metadata and a small sample of values from your warehouse to build the semantic layer. It never copies your full dataset.

Supported sources

• PostgreSQL — live introspection of information_schema plus three sample values per column.
• Demo warehouse — a bundled senior-living dataset (DuckDB) for evaluating Sema with zero setup.
• Redshift / Snowflake / BigQuery / MySQL / REST — on the roadmap; the connector interface is source-agnostic.

Connecting Postgres

1. 1.Open Schema → Connect Postgres.
2. 2.Provide host, port, database, user and password (read-only credentials recommended).
3. 3.Sema introspects the schema and kicks off glossary generation automatically.