One layer of calm
Coverage you can explain
in one breath.
BackupSheep separates what you protect from where you keep it. That simple model makes complex infrastructure easier to operate — and much easier to recover.
FILES / 01
Only move what changed.
Connect any Linux host over FTP, FTPS, SFTP, or SSH. Incremental mode keeps a per-node snapshot cache, so every run is a complete standalone zip without re-downloading the whole site.
Regex + glob include / excludeParallel transfersEd25519 · ECDSA · RSA keys
DATA / 02
Dump it close to the metal.
Use native clients for MySQL, MariaDB, and PostgreSQL. Choose every database or specific tables, connect directly or through an SSH tunnel, and bring back databases that no longer exist.
Version-matched pg_dump 14–18SSL / TLS connectionsStored procedures included
INFRA / 03
Snapshot the infrastructure too.
Protect cloud servers, volumes, managed databases, WordPress, and Basecamp from the same console. Provider-native snapshots stay in your cloud account; offsite dumps and files go to the destinations you choose.
AWS · DigitalOcean · Hetzner · VultrUpCloud · Oracle · Google Cloud · OVHWordPress + Basecamp
Your data, your topology
Choose the place
your recovery lives.
Send each backup to one destination or fan it out across several. Local disk, object storage, or a protected air-gapped copy — keep the policy in one place.
FAN-OUT READY1 source → many destinations
Keep a local working copy, an offsite archive, and an immutable S3 copy in the same run.
Amazon S3Backblaze B2WasabiCloudflare R2DigitalOcean SpacesGoogle Cloud StorageGoogle DriveAzure BlobDropboxOneDrivepCloudIDrive e2IBM COSOracleScalewayLinodeVultrUpCloudExoscaleFilebaseIONOSLeviiaRackCorpTencent COSAlibaba OSSLocal Storage
Immutable S3 archives
Object Lock governance or compliance retention, plus a protected copy your schedule validates first.
Lifecycle controls
Read the lifecycle guide ↗Tier older archives to colder S3 classes and project storage cost by source and destination.
The recovery moment
Make “we need to restore” a button.
Choose a historical backup, choose the copy, and let BackupSheep put the data back. Every restore is tracked with live status and a redacted run log.
- 01Websites
Push files back with an optional exact mirror to remove anything missing from the backup.
- 02Databases
Import native dumps and create databases that no longer exist over direct or SSH-tunnelled connections.
- 03Cloud snapshots
Rebuild through the provider when the backup is a native server or volume snapshot.
RESTORE / WEBSITE ready
production-site
✓Restore all files684 MB
✓Exact mirror delete missing fileson
Notify team when completeoff
Restore is simulated in this preview.
verified✓standalone
archive
archive
Trust is a system property
Quietly serious about
the hard parts.
Backups are only useful if the archive is complete, the credential handling is safe, and the restore works when the clock is loud.
No silent partials.
Every transfer and native command is verified. One failed file fails the run, so it retries instead of archiving a gap.
Preflight the air gap.
Require a protected copy to validate before a schedule runs. No validation, no false confidence.
Credentials stay quiet.
Secrets are encrypted at rest, passed through temporary files, and redacted from run logs.
Day-two operations
Built for teams,
not just backups.
Everything the console does is also available over the REST API, so your backup system can be automated too.
Live dashboard
Storage usage, upcoming runs, and failures that need attention.
Flexible schedules
Daily, weekly, monthly, or cron with keep-last retention.
Notifications
Backup and restore results by email, Slack, or Telegram.
Team accounts
Invite links, groups, permissions, and per-node scoping.
Activity log
Account-wide audit trails, including sign-in tracking.
Scalable workers
Specialized queues you can scale independently as the stack grows.
Self-host anywhere
From zero to protected
in a few minutes.
One command installs the full Docker Compose stack — web console, workers, scheduler, PostgreSQL, and RabbitMQ. Then point it at your own storage.
backup-sheep / installready
# one-command install $ curl -fsSL https://raw.githubusercontent.com/ bilal414/backupsheep/main/install.sh | sudo bash ✓ Docker + Compose installed ✓ app, db, and onboarding secrets generated ✓ health check passed your console is ready →
Ubuntu 22.04+ · Debian 12+self-hosted by design
Short answers
Good questions
welcome.
Is BackupSheep really free?
Yes. It is open source under the GNU GPLv3. You pay only for your own server and storage.
Where do my backups live?
On storage you own — any of the 26 supported destinations or plain disk on the BackupSheep server. Your data does not pass through a BackupSheep cloud.
How do incremental website backups work?
After the first full mirror, only new and changed files cross the network. Every backup remains a complete, standalone archive.
Can backups survive ransomware?
Enable S3 Object Lock and archives cannot be deleted until retention expires. A protected-copy requirement can also stop a schedule before it runs.
What does a restore look like?
Pick a historical backup and click restore. Websites are pushed back, database dumps are imported, and every run keeps live status and logs.