Replication is the replica that keeps running
Standing CDC replicas between SQL Server and Oracle, run by the Executor, with the operating decisions kept in the product.
Until now the honest answer to "can Migratrix keep two databases in sync" was no. Conversion Studio captured changes to shorten a cutover and then stopped. Replication is the same capture turned into a standing replica: pick a SQL Server or Oracle source, a SQL Server or Oracle target, the tables with a key, and the Executor loads them once and then follows the log from a checkpoint it keeps on the target.
What we chose to make visible rather than clever: lag in seconds, applied and rejected rows, a pause with the diff when the source schema changes (you accept it or resync, we do not replay DDL), a consistency check you run on demand, and a Promote step that applies the held-back foreign keys and triggers when the target becomes the system of record.
What it is not: a PostgreSQL, MySQL or MongoDB target, a Kafka feed, or a replacement for a dedicated streaming platform. Keyless tables are refused. It is in the Scale plan.