Supabase Backups: PITR, Restore Steps and Plan Limits

Supabase backups depend on your plan. The Free plan has no automatic backups, Pro keeps 7 days of daily backups, Team 14 days and Enterprise up to 30. Point-in-Time Recovery (PITR) is a paid add-on that restores to the second. Database backups never include the files you keep in Supabase Storage.
Key takeaways
- Free projects have no automatic backups. Export them yourself with
supabase db dumpand keep the files off-site. - Pro keeps 7 days of daily backups, Team 14 days, Enterprise up to 30. A daily backup can still lose up to a day of writes.
- PITR is a paid add-on billed per hour, roughly $100 a month per project for each 7 days of retention, with a worst-case recovery point of two minutes.
- Backups cover the database only. Storage files, Edge Functions, Auth settings and API keys need their own plan.
- Restore to a new project when you can. An in-place restore overwrites production and takes the project offline while it runs.
Plan limits, prices and dashboard paths below were checked against the Supabase documentation and pricing page on November 13, 2026. They change, so each one links its source.
What backups does Supabase include on each plan?
Supabase backs up Pro, Team and Enterprise projects automatically once a day (Supabase: database backups, Supabase pricing):
Supabase's own answer to "does the Free plan get backups" is no: back up manually through the CLI, or upgrade to Pro (Supabase: Free plan backups). One more rule applies on every plan: deleting a project permanently removes its data, including the backups.
What do Supabase backups not include?
A database backup is a copy of Postgres. These parts of a project are outside it:
- Storage files. The database holds only metadata about objects, so restoring an old backup does not restore files deleted after it (Supabase: database backups).
- Passwords for custom roles. Daily backups do not store them. After a restore, reset the password of every custom role you use.
- Project configuration. A restore to a new project copies the schema, data, roles, Auth users and the encryption root key. It does not copy Storage objects and bucket settings, Edge Functions, Auth settings and API keys, Realtime settings, extension settings or read replicas (Supabase: restore to a new project).
The result can be a database full of records whose files are gone. In one audit, a full restore of the database together with file storage had never been demonstrated.
How do you back up a Supabase project on the Free plan?
Use the Supabase CLI to take a logical backup in three files: roles, schema and data. The CLI runs pg_dump in a container, so Docker has to be installed (Supabase: backup and restore using the CLI).
- Click Connect in the dashboard and copy the database connection string. Use the direct connection, or the Session pooler if your network has no IPv6.
- Replace
[YOUR-PASSWORD]with the database password. You can reset it under Database > Settings. - Run the three dumps:
supabase db dump --remote=[CONNECTION_STRING] -f roles.sql --role-only
supabase db dump --remote=[CONNECTION_STRING] -f schema.sql
supabase db dump --remote=[CONNECTION_STRING] -f data.sql --use-copy --data-only \
-x "storage.buckets_vectors" -x "storage.vector_indexes"
Older CLI releases take --db-url instead of --remote; the old flag still works in v3 but is deprecated (Supabase CLI: db dump). Put the commands in a script, run it on a schedule from a machine or CI job you control, add the date to the file names, and copy the files to storage outside Supabase. The files contain your users' data and the role definitions, so keep them encrypted and access-controlled.
Storage files need a separate copy. Supabase publishes a Node.js script that lists every bucket and downloads each file with a secret key; it is in the "Migrating storage objects" section of the same guide. Run it from a server, never from a browser.
How does Supabase PITR work?
Point-in-Time Recovery combines a daily physical backup with archived Write Ahead Log (WAL) files, which record every change the database processes. Supabase uses WAL-G for both. WAL files are uploaded every two minutes by default, or sooner when they pass a size threshold, so the worst-case recovery point objective is two minutes (Supabase: Point-in-Time Recovery).
- Who can use it: Pro, Team and Enterprise projects, as an add-on, on every compute size.
- Where to enable it: the project's Add-ons settings in the dashboard. The Point in Time page under Database > Backups then shows the earliest and latest recovery points.
- What changes: with PITR on, Supabase stops taking daily backups, since PITR covers the same ground at finer granularity.
- What it costs: billed per hour, about $100 a month for 7 days of retention, $200 for 14 days and $400 for 28 days. The add-on is not covered by the Spend Cap (Supabase: PITR usage).
Daily backups fit an app where losing up to a day of writes is acceptable. If a lost day means lost orders, payments or customer records, PITR is the tool that closes that gap.
How do you restore a Supabase project from a backup?
From the dashboard (Supabase: backup and restore process):
- Open Database > Backups. For daily backups, pick the closest backup made before the moment you want to return to. For PITR, click Start a restore and choose a date and time inside the recovery window.
- If the project uses subscriptions or replication slots, drop them first and recreate them afterwards. Supabase handles the Realtime slot itself.
- Confirm. The project is inaccessible during the restore, and downtime grows with database size, so plan for it.
- After the notification that the restore finished, reset passwords for custom roles and test the app.
A restore in place replaces the current database. Anything written after the chosen point is gone, which is why Supabase suggests restoring to a new project first when you want to keep current data.
The same operations are available through the Management API, for example in a runbook script:
# Get your access token from https://supabase.com/dashboard/account/tokens
export SUPABASE_ACCESS_TOKEN="your-access-token"
export PROJECT_REF="your-project-ref"
# List all available backups
curl -H "Authorization: Bearer $SUPABASE_ACCESS_TOKEN" \
"https://api.supabase.com/v1/projects/$PROJECT_REF/database/backups"
# Restore from a PITR backup (Unix timestamp of the restore point)
curl -X POST "https://api.supabase.com/v1/projects/$PROJECT_REF/database/backups/restore-pitr" \
-H "Authorization: Bearer $SUPABASE_ACCESS_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"recovery_time_target_unix": "1735689600"
}'
The commands come from the backups API section of the Supabase docs.
How do you restore to a new project instead of overwriting production?
Go to Database > Backups, open the Restore to a New Project tab, pick a daily backup or, with PITR, a point in time, and click Restore. Supabase creates a new project in the same region with the same compute size, disk, SSL and network settings (Supabase: restore to a new project).
- It is available on paid plans with physical backups enabled, and the new project is billed as a separate project. The cost is shown before you start.
- A project created this way cannot itself be the source of another clone.
- It is a binary copy, so extensions that act on the outside world, such as
pg_cronjobs,pg_netwebhooks and wrappers, start running as soon as the restore completes. If that could send emails or charge cards twice, use a CLI restore instead and inspect those jobs first.
This is also the supported way to get deleted rows back: restore a backup from before the deletion into a new project, then copy the rows you need into production. Supabase warns against recovering rows from dead tuples with tools like pageinspect, which can corrupt data.
How do you restore a CLI dump into a new project?
For the files from the Free plan routine above, create a new project, enable any Database Webhooks and non-default extensions the old one used, and run (Supabase: restore backup using CLI):
psql \
--single-transaction \
--variable ON_ERROR_STOP=1 \
--file roles.sql \
--file schema.sql \
--command 'SET session_replication_role = replica' \
--file data.sql \
--dbname [CONNECTION_STRING]
- Vault and encrypted columns: backup files never contain the root encryption key. Copy it from the old project through the Management API before the old project is paused or deleted, or those secrets stay unreadable.
- Changes to the auth and storage schemas, such as triggers on
auth.usersor RLS policies onstorage.objects, are restored separately.supabase db diff --remoteagainst the old project captures them. - Realtime: re-enable publications under Database > Publications.
How often should you test a Supabase restore?
Often enough that you know the steps and the duration before you need them, and always after a large schema change. A restore drill takes an afternoon:
- Write down how much data you can afford to lose and how long the app can be down. Those two numbers decide between daily backups and PITR.
- Restore the latest backup to a new project, never over production.
- Restore the matching Storage files, point a copy of the app at the new project, and check that records and their files both open.
- Record how long each step took, and keep the notes with the runbook.
Backups and restore drills are standard items in DevOps best practices, and investors ask about them in technical due diligence.
What good looks like: Supabase backup checklist
What good looks like
- You know which plan the project is on and how far back its backups reach.
- Free projects have a scheduled
supabase db dumpjob with copies stored outside Supabase. - Production projects that cannot lose a day of writes run PITR.
- Storage files are backed up separately from the database.
- A restore to a new project has been done, with records and files checked, and its duration written down.
- Custom role passwords, Vault keys and cron or webhook jobs are covered in the restore runbook.
Backups are one line of the production readiness checklist for vibe-coded apps.
Not sure your app would survive a bad deploy?
Valletta Software runs a fixed-price vibe code audit, from $199, for apps built with AI tools. We read your source code, including migrations and policies, and we never touch production or customer data. The audit covers backups and restore readiness along with security, data handling and code quality.
Need someone to set up backups, PITR and restore drills for you? See DevOps support.
Supabase backups FAQ
Does Supabase back up projects on the Free plan?
No. Supabase makes automatic daily backups only for Pro, Team and Enterprise projects. On the Free plan, Supabase recommends exporting your data regularly with the CLI command supabase db dump and keeping the files off-site. Upgrading to Pro turns on daily backups with 7 days of retention.
How long does Supabase keep backups?
Daily backups are kept for 7 days on Pro, 14 days on Team and up to 30 days on Enterprise, according to the Supabase backups documentation. With Point-in-Time Recovery the window is the retention period you pay for, such as 7, 14 or 28 days. Deleting a project removes its backups permanently.
What is Supabase PITR and how much does it cost?
Point-in-Time Recovery restores your database to any second inside its retention window by replaying archived write-ahead logs on top of a physical backup. It is an add-on for paid plans, billed hourly: about $100 a month for 7 days of retention, $200 for 14 and $400 for 28, per project.
Do Supabase backups include Storage files?
No. Database backups contain only the metadata about Storage objects, not the files. Restoring an old backup does not bring back files deleted after it. Restore to a new project does not copy Storage objects or bucket settings either, so back up important files separately and test that records and files match after a restore.
Can I download a Supabase backup?
Projects on physical backups or PITR do not get a downloadable logical backup file. To get a copy you can keep, run supabase db dump or pg_dump against your database: one dump for roles, one for the schema and one for the data. Store the files outside Supabase and test that they restore.
How do I restore a Supabase backup without losing current data?
Use Restore to a New Project, under Database > Backups. It creates a separate project from a daily backup or a PITR point and leaves production untouched. You can then copy the rows you need back. Restoring in place replaces the current database and makes the project unavailable while it runs.