Cloud backups

Cloud can back up a managed WordPress, PrestaShop or PRISM site to archive storage. Backups are tied to the site's owning account and the daily-backup subscription option.

Check availability#

Open your project's Backups section. It shows whether backups are enabled and lists recent runs. If the option is unavailable, review it from your account's subscription and billing page.

This backup workflow applies to Cloud-managed sites. An arbitrary project or externally deployed application does not automatically become eligible because it has a project ID. Administrators also need to use the owning account's backup workflow; the backup API does not grant access to another user's archives.

Create a backup#

Choose Back up now on the project page, or ask a connected assistant to call create_backup with the project ID.

The API equivalent is:

bash

curl -X POST https://cloud.bext.dev/api/backups \
  -H "Authorization: Bearer $BEXT_CLOUD_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"project_id":"YOUR_PROJECT_ID"}'

A successful response means Cloud accepted the backup request. The engine creates the archive in the background. Poll the history before attempting to download it:

bash

curl "https://cloud.bext.dev/api/backups?project_id=$PROJECT_ID" \
  -H "Authorization: Bearer $BEXT_CLOUD_TOKEN"

State	Meaning	Next step
`running`	The archive is still being created or uploaded	Wait and refresh the history
`done`	The run completed	Request a private download link
`failed`	The run did not complete	Read its error before retrying

An active run blocks another backup of the same site. Manual backups also have a 15-minute cooldown after a recent successful or running backup. Retrying a refused request repeatedly will not bypass it.

Download an archive#

Use the download control for a completed run on the project page. Through MCP, call download_backup with project_id and backup_id.

For the API, take the backup ID from the history response:

bash

curl "https://cloud.bext.dev/api/backups?project_id=$PROJECT_ID&download_id=$BACKUP_ID" \
  -H "Authorization: Bearer $BEXT_CLOUD_TOKEN"

The response contains a private URL valid for 15 minutes. Download the archive before it expires. If the link expires, request another one for an archive that is still retained. Anyone with that URL can use it while it is valid, so share it only with the owner or an authorized recovery operator.

The history endpoint and list_backups tool do not include download URLs. A download URL is issued only when you explicitly request one.

Scheduling and retention#

With the daily-backup option enabled, an hourly scheduler checks which managed sites are due for a backup. This is a daily backup workflow, not a guarantee that a run begins at an exact wall-clock time.

Archives are retained for 30 days. Retention cleanup also applies after the option is disabled; switching it off does not preserve old archives indefinitely. Keep an independent downloaded copy if you need a longer retention period.

Restoration limits#

Cloud does not currently expose self-service backup restore or manual archive deletion through the dashboard backup workflow or its MCP tools. A download is an archive, not a restoration job.

For recovery help, use support and identify the site and backup run without posting private URLs or account tokens. Before a live restoration, establish which files and database will be replaced and retain the current state as a recovery point.