Migrate from Railway

Get the source connection string

In the Railway dashboard: open the Postgres service → Variables tab. Railway exposes two URLs:

• DATABASE_PUBLIC_URL — routed through the TCP proxy (*.proxy.rlwy.net:<port>), reachable from the internet. This is the one you want.
• DATABASE_URL — the private-network URL (postgres.railway.internal), only reachable from inside Railway. CapyDB cannot use it.

postgresql://postgres:password@yamanote.proxy.rlwy.net:12345/railway

Railway-specific caveats

• The public URL goes through Railway's TCP proxy and bills egress; a one-time logical copy is normally trivial, but a very large database costs what it costs.
• If the service has TCP proxy disabled (private-only networking), enable a public endpoint temporarily for the import, then disable it again.
• The default database is literally named railway — the import recreates the CapyDB project database under its own generated name; your app does not care, the URL carries the name.

Preflight

capydb import preflight --source-url "postgresql://postgres:password@....proxy.rlwy.net:12345/railway"

Railway Postgres images are recent majors with few default extensions, so the version and extension checks usually pass; size against the plan is the one to watch.

Import

Pause writers, then:

capydb import --source-url "postgresql://...proxy.rlwy.net:12345/railway" --recreate --wait

After the import

1. Spot-check sequences.
2. \dx for extensions.
3. ANALYZE;
4. Cut over env vars in your Railway app services (or wherever the app now lives), deploy, and watch Observability.