Restore a Hyper Backup without a Synology NAS

This guide is for you if your NAS is gone and the backup is all you have. The backup is a folder whose name ends in .hbk, on a USB disk, a NAS share or a cloud bucket. The guide is also for you if you want to prove that your backup restores before you need it.

If the disk is slow, clicks or shows SMART errors, stop. Image it or send it to a recovery lab, then run hbkit on the copy.

hbkit never writes to the archive. If it is your only copy, copy the whole .hbk folder to another disk first and run hbkit on the copy.

hbkit is an open reader for your own backups. It is free and open source (MIT), and reads .hbk archives directly, with no Synology software. It has restored a 3 TB archive byte-exact. Its doctor command checks the layout and rebuilds a sample of files before you restore anything:

$ hbk demo/synth.hbk doctor
          chunk codec : lz4
            encrypted : no
               shares : Documents, Photos
      backup versions : 3
      reading version : 3 (2026-10-28 08:33:20, Complete) - newest complete

  PASS  virtual_file layout known  (56 B)
  PASS  chunk_index layout known  (29 B)
  PASS  bucket layout known  (32 B)
  PASS  rebuilt 6 sampled files, all chunks verified  (6 ok, 0 failed)

  VERDICT: recoverable. Sampled files rebuilt byte-exact and checksum-verified.

When the answer is no, the verdict says so: "cannot be recovered by this tool" or "do not trust output". After you install it, hbkit works offline for a local disk and sends nothing anywhere. It never writes to the archive.

What hbkit can't do yet

• Synology C2 Storage: Not tested yet. Use Hyper Backup Explorer or C2 Storage Explorer in the C2 web portal.
• Entire System backups and DSM settings: hbkit restores files and folders. To put a whole system or its settings back, you need a Synology NAS.
• Damaged or encrypted files in Config/, for example after ransomware: hbkit needs the file index in Config/ and stops without it.
• Key-file access: not supported. Use the password. Hyper Backup Explorer is the right tool if you only have the key file.
• Older RSA-encrypted archives: not supported. Use Hyper Backup Explorer.
• Restore on Windows: On Windows, CI restores test archives byte-exact on every change. A backup from a real NAS has not been restored on Windows yet.

Five steps

1. Copy the whole .hbk folder first

The backup is a folder such as nas_1.hbk, with Config/ and Pool/ inside. The file SynologyHyperBackup.bkpi is only a marker, and it can be empty. Your data is in Pool/. The file names and folders are in Config/. If you copy the backup, copy the whole .hbk folder, not only the .bkpi file.

hbkit never writes to the archive. If it is your only copy, copy the whole .hbk folder to another disk first and run hbkit on the copy. For a bucket, a copy on a local disk is much faster.

2. Install hbkit

brew install YordiLorenzo/tap/hbkit

Homebrew works on macOS and Linux, and it installs liblz4 for you. Without Homebrew, use pipx (or pip) and install liblz4 yourself:

pipx install hbkit
sudo apt install liblz4-1      # or: brew install lz4, sudo pacman -S lz4

hbkit needs Python 3.10 or newer.

3. Check the archive with doctor

hbk /Volumes/Backup doctor

Give it the .hbk folder, or any disk or folder that holds one. doctor reads the layout and rebuilds a sample of real files. It checks each of their chunks against the MD5 and CRC32 stored in the archive. For an encrypted archive, add -p and the password or set HBK_PASSWORD. doctor does not prompt for a password.

If the verdict is not "recoverable", do not restore yet. Open an issue and paste the output. The fixes in 0.4.5 and 0.4.6 came from reports like that (#1, #4).

4. Pick the backup version

hbk /Volumes/Backup versions

This lists every backup version with its start time and status, and marks the default. It needs no password. hbkit reads the newest complete version unless you choose another. --version 12 reads version 12. --version @2026-10-15 reads the newest complete version that started by the end of that day:

hbk /Volumes/Backup list invoice --version @2026-10-15

5. Restore

To tick folders in a full-screen browser:

hbk-tui /Volumes/Backup

Press space to tick a folder, d to choose the destination and r to start the restore. Before you start, it warns you if the selection will not fit. Or use one command:

hbk /Volumes/Backup get "/Photos/*" ~/restore

Paths start with a slash and the share name. Files go to ~/restore/Photos/... with their original dates. If the restore stops, run the same command again. It skips the files that are complete.

Where the backup is stored

Ways to get the files back

hbkit is not always the right tool.

The README names two more open source readers under Prior art.

Linux, without a desktop

Synology's Explorer for Linux is a desktop app for x86-64 only. On a server, use hbkit over SSH:

pipx install hbkit
sudo apt install liblz4-1
hbk /mnt/backup doctor
hbk /mnt/backup get "/Documents/*" /srv/restore

hbk-tui runs in the terminal, so it also works over SSH. For a restore test from cron, doctor exits with 0 when the archive is recoverable and 1 when it is not. verify checks files without restoring them. It can write a local index cache, but never changes the archive. The README lists every exit code.

Restore from a bucket

Mount the bucket with rclone and point hbkit at the mount. hbk mount sets the rclone flags that matter and loads the folder listings once, at the start. It stores no keys. rclone config keeps them.

From R2, a restore ran at about 9.5 MB/s with -j 16. Decompressing and checking took 0.7% of the time, in that measurement. From Backblaze B2, a contributor rebuilt a 74,396,840-byte video exactly, with 9,072 chunks checked (PR #3). Start with one small folder and watch the speed.

Encrypted archives

hbkit reads archives with Hyper Backup's client-side encryption, with the password. It derives the public key from the password and compares it with the key in the archive. A wrong password is rejected before any data is read.

• versions needs no password, because the version list is not encrypted.
• The index cache in ~/.cache/hbkit holds the decrypted file names, not the contents. Delete it after the restore if that matters to you.
• Tested on a small encrypted archive from Hyper Backup 4.1.2 and a 7-version encrypted archive from Hyper Backup 4.2.2 (#4).
• Not supported: access with the key file, and older RSA-encrypted archives. Use Hyper Backup Explorer for those.

Leaving Synology

Synology's spec page says multi-version backups "can be read only with Hyper Backup, Hyper Backup Explorer, and Hyper Backup Vault." All three are Synology software. hbkit reads these backups without Synology software, and FORMAT.md is a public description of the .hbk format, written so archives stay readable. The Mac version of Explorer needs Rosetta. hbkit runs natively on Apple silicon. See the Mac guide for setup and ext4 USB disks. Before you move to other hardware:

1. Run doctor on each archive from a computer, not from the NAS.
2. Restore the data you want to keep into plain folders.
3. Keep the .hbk folders. hbkit can still read their older versions with --version.

For errors during a restore, see damaged backups.

Start with doctor

brew install YordiLorenzo/tap/hbkit
hbk /Volumes/Backup doctor

doctor writes nothing to the backup. If it does not say "recoverable", open an issue with its output. Every command is in the README.