Managed PostgreSQL

Every container project (not static sites) that isn't over its plan's database quota gets a dedicated PostgreSQL database and role, provisioned automatically the first time the project is created — you don't request it separately.

What you get

• One database, one role that owns it, one randomly generated password — isolated from every other project on the platform, including other projects in your own account.
• A connection is capped at a fixed CONNECTION LIMIT set on the database itself (currently 10 on every plan) — this is enforced by Postgres, not just advisory, so a connection-leaking app gets a hard too many connections error rather than starving other tenants on the shared cluster.
• The connection string is injected into your app automatically — see Connect to DATABASE_URL.

The database is only dropped when the project is fully purged (not just deleted). A soft-deleted project keeps its database until an operator releases the resources — this is intentional, so an accidental delete is recoverable.

Table Editor and SQL Editor

The project's Database page lists your tables (with column types, row counts, and size) and lets you browse, filter, and inline-edit rows without leaving the dashboard. The SQL Editor runs arbitrary SQL directly against your own database — same trust level as connecting with psql, so it isn't sandboxed or restricted to SELECT.

Schema editing, CSV, and saved queries

• Create/drop tables and columns, and create/drop indexes, from the Schema tab on a table's page — no SQL required, though the SQL Editor can do the same thing with CREATE/ALTER/DROP.
• Export CSV downloads up to 50,000 rows of a table. Import CSV reads a file's header row as column names and batch-inserts matching rows — columns that don't exist on the table are ignored, and empty cells become NULL.
• The SQL Editor can save a query by name for later reuse — saved queries are scoped to the project and stored in AnySites' own database, not yours.

ER diagram, backup, and restore

• The ER Diagram page draws every table and foreign-key relationship from the live schema — nothing to configure, it's generated on every visit.
• Backup & Restore downloads a full pg_dump (custom format) of the database on demand, and can restore one back — restore runs pg_restore --clean, which drops and recreates everything in the file, replacing the current contents entirely. There's no automated/scheduled backup yet; download one yourself before a risky change.

Related documentation