Choose CLI or Server

ReplicaDB keeps the direct transfer tool and the managed control plane as separate products. Choose based on who owns state, scheduling, and access.

Standalone CLI

Choose the CLI when a script, pipeline, or operator owns each replication invocation. It runs on Java 17 or newer, accepts command-line options or an options file, and does not require ReplicaDB metadata storage.

The CLI is a good fit when you need:

• a direct source-to-sink transfer from a terminal or scheduler;
• an existing secrets and logging workflow around the process; or
• the smallest operational footprint for one host.

Start the CLI quickstart.

Managed Server

Choose the managed server when replication needs durable jobs, schedules, authenticated access, run history, diagnostics, or multiple operators. The server uses PostgreSQL metadata and exposes a web control plane; local mode manages embedded PostgreSQL, while api and worker deployments use external PostgreSQL.

The server is a good fit when you need:

• reusable datasource profiles and jobs;
• permissions, audit history, and shared visibility; or
• local or distributed execution with durable recovery state.

Start the server quickstart.

They are related, not interchangeable

Both products use the ReplicaDB replication core and support the same broad connector family, subject to connector-specific limits. A server installation does not move CLI state, and a CLI invocation does not create managed jobs. Keep REPLICADB_HOME for CLI installations and REPLICADB_SERVER_HOME for server installations.