Getting Started

agent setup

Using an AI coding agent?

Copy the setup prompt and paste it into your agent from inside your project. It reads your model, and wires Wallaby in.

Server prerequisites

Your Postgres 15+ server must already have:

wal_level = logical set in postgresql.conf - required for logical replication. (Not required for PG 19).
A role with the REPLICATION attribute (or superuser) for the connection string you give Wallaby.
Headroom in max_replication_slots and max_wal_senders (at least one slot/sender per Wallaby cluster).

Wallaby validates these on startup - including the server version - and fails fast with an actionable error if something is missing.

Choose your configuration

Wallaby is configured through AddWallaby, with a storage provider supplying the model of what to capture and a sink as an output destination. Pick the setup that fits your application. Optionally, if you don't want Wallaby to capture anything itself, let it provision and maintain publications and replication slots for an external pgoutput consumer:

Mixing Providers

The EF Core, Marten and plain-table providers can run side by side in one Wallaby instance, sharing a single replication slot - see the providers overview.

Deployment

It's highly recommended to deploy Wallaby as a separate service, not as an integrated part of your main application. This allows you to scale CDC operations independently as the need arises.

Next steps

How it works: the capture pipeline end to end.
Configuration: all configuration options.
Mappings: routing entities to destinations, shaping and enriching documents.
Backfill: initial snapshots and version-triggered reindex.
Observability: OpenTelemetry metrics and traces.

Getting Started ​

Server prerequisites ​

Choose your configuration ​

Deployment ​

Next steps ​

Getting Started

Server prerequisites

Choose your configuration

Deployment

Next steps