이동성과 이전

Chapter VI of Regulation (EU) 2023/2854 - the Data Act - has applied since 12 September 2025 and covers data-processing services including managed databases. It requires providers to publish how a customer can switch away, what can be ported, how long it takes, and what it costs. This page is that publication, and the corresponding commitments are part of the Terms of Service.

The short answer: everything is exportable with standard PostgreSQL tools, there is no proprietary format to escape from, and we do not charge you to leave.

• -Complete database contents - schema, data, indexes, constraints, functions, and extensions - as a pg_dump custom-format archive, either self-serve or through a managed export job
• -A live logical replication stream, so you can cut over with near-zero downtime instead of taking an outage
• -Schema-only dumps, plus generated TypeScript, Zod, and Drizzle definitions of your schema
• -Point-in-time backup contents, restorable into a cell you can then dump
• -Project and workspace metadata: projects, regions, plans, connection topology, preview databases, backup schedules, audit entries, and webhook endpoints, through the API
• -Your account and organization records, on request

Your connection string is a normal PostgreSQL URL, so every standard tool works without modification. Nothing here requires our cooperation, our permission, or a support ticket.

For a live migration out, point your target's logical replication or migration tool at the cell the same way you would at any self-hosted Postgres. There is no gateway in the path that will refuse it.

# Full logical dump with standard tooling
pg_dump "$DATABASE_URL" -Fc -f capydb-export.dump

# Schema only
capydb schema dump --out schema.sql

# Managed export job, queued, awaited, and downloaded
capydb export --project my-project --output capydb-export.dump

We will help you leave: we answer questions about the export, help you size the target, and stay on the thread through the cutover. That assistance is free and is not conditioned on telling us why you are going.

If an export exceeds what a single dump can comfortably carry, we will produce a staged export and coordinate the timing with you.

Nothing. CapyDB charges no switching fee, no egress fee, no early-termination charge, and no per-gigabyte export charge. This is our position today, not a reduced charge under a transitional allowance.

For context: the Data Act permits reduced switching charges only until 12 January 2027, after which all switching charges must be eliminated. We are already at zero and will stay there.

You continue to pay for the subscription itself for as long as the project exists, because the databases are still running.

• -Self-serve export: immediate. A dump starts the moment you run it, and finishes at the speed of your data.
• -Assisted switching: we respond to a switching request within one working day and agree a plan with you.
• -Transition period: up to 30 days from the request, extendable where a migration is genuinely technically infeasible in that window, in which case we tell you why and when.
• -Deletion after switching: on your confirmation, cells are destroyed immediately and their backup copies age out of the archive within the retention window of your plan.

CapyDB runs unmodified PostgreSQL. Your schema, your queries, your extensions, and your client libraries all work the same on any other PostgreSQL - managed, self-hosted, or on a laptop - which is the point of building the product this way. There is no proprietary query API, no bespoke driver, and no data model to translate.

The only CapyDB-specific objects in your database are the platform helpers the product installs, and they live in their own schema so a dump of your own schema does not carry them.

The same standards work inbound. We import from other managed providers and from self-hosted Postgres, including a follow mode that keeps the source and the cell in sync until you cut over. Migration guides and codemods for the common origins are in the documentation.