# Multigres

Horizontally scalable Postgres for high availability

Run high availability Postgres across multiple nodes with automatic failover, using Multigres.

Caution: Multigres is in [Public Alpha](https://supabase.com/docs/guides/getting-started/features#feature-status). It's free during the Alpha for organizations on a paid plan, for up to two projects, but pricing may change once it leaves Alpha. It isn't covered by the [uptime SLA](https://supabase.com/sla), and Supabase isn't targeting production or mission-critical workloads with it during this stage.

Multigres gives your project high availability by running Postgres across multiple nodes instead of one, so reads and writes continue if a node fails. It is Supabase's integration of [Multigres](https://multigres.com), an open-source project that brings the same distributed-systems approach Vitess brought to MySQL to Postgres.

## What you get

When you enable Multigres on a project, your database runs as a small cluster instead of a single instance:

- **Automatic failover:** If a node fails, another in the cluster is promoted within seconds, without you having to intervene.
- **No connection changes:** Use the same connection string. Coordination is transparent to your application.
- **Consensus-backed durability:** Writes are acknowledged only after the cluster agrees they are durable.

## Eligibility

During the Public Alpha:

- Multigres is available to organizations on a paid plan. It isn't available on the Free plan.
- Availability is being rolled out gradually, so the option to enable it may not yet appear for every eligible organization.
- Up to two projects per eligible organization can use Multigres for free during the Alpha.

## Enabling Multigres

In the Dashboard, Multigres appears as **High availability** during project creation:

1. Open [Create a new project](https://supabase.com/dashboard/new/_) in the Dashboard.
2. Under **High availability**, turn on **Enable high availability**.
3. Finish the remaining fields such as database password, region, and compute size, if shown.
4. Click **Create new project**.

If your organization isn't eligible or hasn't been rolled out yet, **High availability** won't appear.

## What's not included in the alpha

Some Supabase features and project operations aren't yet available on Multigres-backed projects:

- **Realtime.** Multigres doesn't yet support the logical replication that Realtime depends on, so Realtime is unavailable.
- **Point-in-Time Recovery.** Only the standard daily backups are available.
- **Cross-region read replicas.** Multigres projects run in a single region during the Alpha.
- **OrioleDB.** A project can use Multigres or [OrioleDB](https://supabase.com/docs/guides/database/orioledb), not both.
- **Resizing after creation.** Compute size, disk, and the number of replicas can't be changed after a Multigres project is created.
- **Sharding and custom durability policies.** These are on the longer-term roadmap but aren't part of the Alpha.

Caution: Enabling Multigres migrates your project's database to run on Multigres. There's currently no managed path to move it back to a standard Postgres project.

## Compatibility

Multigres aims for full compatibility with standard Postgres, but there are some differences to be aware of. See [Multigres compatibility](https://supabase.com/docs/guides/database/multigres/compatibility) for details.

## Resources

[Multigres documentation](https://multigres.com/docs) — architecture, self-hosted deployment, and other technical depth that goes beyond the hosted Supabase integration covered on this page.