list_database_backups

These are the archives the Database tab of Backups lists. Each one is a .zip holding database.sql, a plain SQL dump of every table and every row. A row carries its _id, its name, who took it, when, how large the archive is, how large the database was and how many tables and rows it held.

A row typed manual was taken deliberately, and automatic is the daily backup the platform takes on a paid plan, only while daily database backups are on for the site and only on a day the data actually changed. A row reads pending while it is still being exported, ready once it is finished, and failed with the reason beside it when it stopped.

eligibility reads ready when the site can be backed up at all, unsupported when the site is not on the newer database, and no-database when it has no database yet. On anything but ready no backup can be taken and none will ever appear. canBackup is false whenever eligibility is not ready, and also when the workspace is not on a paid plan or the plan could not be confirmed, so read eligibility first and treat canBackup as an answer about the plan only while eligibility reads ready.

planUnknown tells those last two apart: when it is true, the tool tells the client to say your plan could not be confirmed and nothing about upgrading, because create_database_backup is refused for that same reason until it can be. The response also carries the manual allowance in quota, as limit, used and remaining, counted apart from storage backups, the total count, and dailyEnabled, which is false when daily database backups were turned off with set_database_backup_schedule.

Use a row's _id as the backupId for delete_database_backup, and pass page to go further back, since it returns 10 per page. A database backup has nothing to browse and cannot be restored, through MCP, the HTTP API or by anyone: rows come back only when you download the archive in the editor and put the data back deliberately. Older archives are thinned as they age, and nothing survives past 1 year except the newest finished one.

Inputs

Input	Type	Required	Description
projectId	string	Yes	The site id, from list_sites.
page	integer	No	Which page of the list to read, starting at 1. Defaults to 1.