DbDataSync 2026.9.25.1104

DbDataSync

DbDataSync · Install · Configuration · Getting started · Replication concepts · Drivers and libraries · State database · Building from source

DbDataSync is a cross-database replication tool. You define a replication in a web UI — the source table, the target table, column mappings, a schedule, and how changes are processed — and DbDataSync keeps the target in sync. It does a full initial load, then applies incremental changes on a schedule or on demand.

A replication's source and target can be different engines — a SQL Server source can feed a PostgreSQL target, for example. Support varies by engine:

Every engine above also offers TriggerAudit, a portable trigger-based option — see Replication concepts. DuckDB is a source only, for query-based sources like Parquet, CSV, an S3 glob, or an attached database. It is not a replication target. See architecture/planning/done/overview.md for the broader ambition and architecture/detailed-design.md for the full system design.

The table above is the five built-in drivers. Anything else reachable through an ADO.NET provider — SQLite, Firebird, anything ODBC — plugs in as a source without a DbDataSync rebuild, through a YAML descriptor. See Drivers and libraries.

Install

dotnet tool install -g DbDataSync
dbdatasync setup

This installs DbDataSync into your own profile and walks you through setup. See Install for a Windows or systemd service, a machine-wide install, or running in a container. The package is also on nuget.org.

Next steps

• Configuration — every flag and environment variable.
• Getting started — set up your first replication, with screenshots.
• Replication concepts — change detection, bulk loading, and delete reconciliation.
• Drivers and libraries — add another engine (MySQL, Oracle, ...) without a rebuild.
• State database — what DbDataSync tracks about its own runs, and where it lives.
• Building from source — the dev loop, tests, and repository layout.