Your first pipeline

1. Settings → Runtimes: add a Kubernetes namespace or Docker host (or install an agent).
2. Connections: add the source. BeziSync runs readiness checks (logical replication, privileges, retention, keys…) and shows each failure with the exact fix for your provider (self-hosted, RDS, Cloud SQL, Azure…) and a "send to my DBA" message. Add the destination the same way.
3. New pipeline: pick the source (or start from a template), select tables and choose what happens to personal data (see Personal data), pick the destination and naming, shape the events (optional, see below), review the generated application.properties, and deploy. A step opens once the ones before it are done, and leaving a step by any route saves it. On Review, settings save as you change them and the final checks re-run on their own; Deploy says why while it is unavailable. Final checks must pass; "Deploy anyway" (if the project allows it) needs a recorded reason and never skips configuration, runtime or policy failures. Closing a fresh draft asks whether to keep or delete it.