Supabase Pipelines is a managed change data capture (CDC) product that uses Postgres logical replication to deliver published data to analytical destinations in near real time.
Key benefits
- Initial sync: Copy existing rows from published tables.
- Ongoing replication: Apply subsequent INSERT, UPDATE, DELETE, and TRUNCATE operations selected by the publication.
- Managed operation: Monitor pipeline status, lag, table state, and errors in the Dashboard.
- Workload isolation: Keep analytical queries away from the primary database.
Destinations
BigQuery is currently available. Request early access to ClickHouse, Snowflake, and DuckLake while destination support expands.
Setup
Create a Postgres publication for the tables to replicate. In Database > Replication, add a Pipelines destination, configure its settings, and monitor the pipeline from the Dashboard.
Requirements
Requirements depend on the destination. BigQuery requires source tables to have primary keys and requires the publication to include those columns.
Pipelines is valuable for:
- Near real-time analytics data movement
- Analytics separation from production
- Managed replication to supported destination systems
Limitations
Schema change support is currently in beta and limited to supported BigQuery changes. Destination-specific constraints apply.
Pipelines keeps the current destination table state synchronized. It does not automatically create a queryable history of every row version.