Export a Supabase database with pg_dump

Pick the right connection string

Supabase recommends the direct connection for pg_dump, backup, restore and replication. The shared pooler is IPv4-only, which is why the session pooler on 5432 is the usual choice from a laptop or CI runner without IPv6.

Sources: Supabase: Connect to your database, Supabase: Backup and restore using the CLI

Recommended: supabase db dump (three files)

The CLI runs pg_dump in a container with the matching version and skips the schemas Supabase manages, so the files restore cleanly into another Supabase-compatible project.

export DB_URL="postgresql://postgres.<ref>:<password>@aws-0-<region>.pooler.supabase.com:5432/postgres"

supabase db dump --db-url "$DB_URL" -f roles.sql --role-only
supabase db dump --db-url "$DB_URL" -f schema.sql
supabase db dump --db-url "$DB_URL" -f data.sql --use-copy --data-only \
  -x "storage.buckets_vectors" -x "storage.vector_indexes"

psql --single-transaction --variable ON_ERROR_STOP=1 \
  --file roles.sql --file schema.sql \
  --command 'SET session_replication_role = replica' \
  --file data.sql --dbname "$TARGET_DB_URL"

Sources: Supabase: Backup and restore using the CLI

Plain pg_dump (when you do not want the CLI)

A single custom-format archive is handy for a local copy or a non-Supabase Postgres. Your pg_dump major version must be at least the server's; check the server with select version(); and the client with pg_dump --version.

• "server version mismatch": install a newer client (for example brew install postgresql@17) and call that pg_dump explicitly.
• "prepared statement ... already exists" or a hang: you are on port 6543. Switch to 5432.
• "Tenant or user not found": on the pooler the user must be postgres.<ref>, not postgres.
• "Network is unreachable" on db.<ref>.supabase.co: that host is IPv6. Use the session pooler string instead.
• Storage files are not in the dump. Database backups only hold their metadata; copy buckets separately.

pg_dump "$DB_URL" -Fc -v --no-owner --no-privileges \
  --schema=public -f supabase-public.dump

# restore into any Postgres
pg_restore -v --no-owner --no-privileges -d "$TARGET_DB_URL" supabase-public.dump

Sources: Supabase: Connect to your database, Supabase: Database backups

If the project is paused

A paused project refuses connections, so pg_dump cannot reach it. Resume it in the dashboard, or download the backup file from the paused project page.

Sources: Supabase: Project pausing

Moving the dump to Scribase

Scribase imports a Supabase project straight from the same connection string: schema, grants, auth users with their password hashes, data, RLS policies and Storage objects, ending with row counts and checksums per table. Your supabase-js code keeps working with a new URL and key.

Each project is $25 a month, flat, and has an export button that produces the schema, data, auth users and storage in one bundle. Direct psql connections into Scribase are not open yet, so the importer is the way in today.

scribase import supabase --project-ref <ref> \
  --org acme --project app --env production \
  --connection-string "$DB_URL" --dry-run

Sources: Scribase pricing