On this page
1. Prepare the folder and its destination #
Start with a small folder of documents or website assets. Close any application actively rewriting those files. For an application with its own export process, produce that export first; our Thunderbird backup guide illustrates why the recoverable application data matters more than a few visible documents.
Create a destination directory dedicated to this backup. For example, a folder called Project-backup on an external drive should be separate from the project's working directory. Every Duplicati job needs its own destination folder, according to the destination documentation. Two jobs writing into the same folder can interfere with each other's backup data.
Open Duplicati. If its interface isn't visible, choose Open from the tray icon. You should see its local management interface. The free client can work independently of a hosted console account.
2. Back up a folder with Duplicati using a new encrypted job #
Select Add backup, then Add a new backup. Give the job a recognizable name. Choose an encryption method and passphrase, then continue through the wizard with Continue. These labels follow Duplicati's current setup documentation.
Save the passphrase somewhere you can access after this computer fails. A password manager can hold recovery information, but plan how you'll access that manager without the backed-up machine. Duplicati cannot recover an encrypted backup if you lose its passphrase.
The built-in AES method doesn't require installing GPG separately. Duplicati's encryption documentation explains the available methods. For this first backup, I'd use the built-in method and concentrate on completing the recovery exercise.
3. Select and check the storage destination #
Choose the destination type and enter its details. A file destination can be a path available to the Duplicati process, including local or mounted storage. A cloud or SFTP destination needs its own account and connection details.
Use Test destination in the current backup-setup flow. A successful check should confirm that Duplicati can communicate with the chosen destination. Then select Continue.
Check the entire path, especially with network storage. A drive letter visible in your signed-in Windows session might not exist for a separately running service. Make sure the account running Duplicati can reach and write to the directory. The connection check is the point to resolve that problem, before selecting a large source.
4. Select the source and schedule #
Use the source picker to select your working folder. Duplicati includes its subfolders; deselect anything you want excluded. Select Continue, choose the schedule, and continue again.
For a first project backup, select the actual folder containing your work. If it contains several similarly named export directories, check which one is current. An elegant schedule won't help if it backs up an empty staging directory.
Set the run time for when both the computer and destination are available. A removable drive that lives in a drawer won't be reachable during an automatic job. After setup, the practical check is whether a later scheduled run appears in the job's history without manual intervention.
5. Set retention and run the backup #
Choose retention in the final setup screen. Smart backup retention keeps a selection of daily and older versions. Select Save. The new job should appear in the backup list.
Start it with Run now, the action described in Duplicati's graphical-interface documentation. Watch the progress and wait for completion. The initial backup processes all selected data; later runs process new or changed data.
Choose retention by how far back you might need to recover. If a mistake takes weeks to discover, a tiny number of recent versions can be inadequate even when every daily run succeeds. Also allow space for old versions of large files. Retention limits the history you can select during the next step.
6. Restore a copy into a separate folder #
Open Restore and select the configured backup. Choose its backup version and the files or folder you want, then select Continue. Choose another restore destination, such as an empty Project-restore-check directory. Select Submit to start recovery. The restore documentation describes this flow and its summary screen.
Open recovered files with their normal applications. For a website project, inspect a source file and an image, and confirm that expected subfolders came back. Keep the working folder intact during this check. A restored file that only looks correct in the backup browser hasn't yet proved that the application can open its contents.
Record the version you selected and the restore location. Those two details help when a later recovery returns an older file than expected. If an application needs several files together, recover that set rather than one isolated file from the middle of it.
7. Export the job's recovery settings #
Expand the backup configuration and select Export .... Choose To File. If you leave Export passwords unchecked, save destination credentials and the backup passphrase separately. If you include secrets, encrypt and protect the exported configuration. Duplicati's export documentation explains both options.
Store the configuration where a replacement computer can access it. Keeping it only inside the same working folder makes the recovery process depend on the very backup you're trying to reopen. An encrypted export also has its own file passphrase, which can differ from the backup passphrase.
Known problems reported by users #
A NAS installation can't see the computer's files
In a March 2025 forum thread, someone installed Duplicati on a NAS and expected it to reach their PC Desktop. Developer Kenneth Skovhede explained that the program generally needs to run where the source files are available. For this guide, install the client on the computer holding the folder, then use the NAS as its destination.
Restored files have the wrong access permissions
A December 2023 discussion about restoring to another user explains how saved numeric user and group IDs can differ on the replacement system. Leave permission restoration disabled when recovering personal documents to your own empty folder. Duplicati's current restore documentation also warns that restoring old permissions can make the recovered files inaccessible to the new user.
Troubleshooting the backup and restore #
| Problem | Likely cause | What to do |
|---|---|---|
| Destination check fails | Incorrect path, credentials, or write access | Confirm the destination using the same account that runs Duplicati; repeat Test destination |
| Source folder is missing | The client or service cannot see the path | Run the client where the source is available and check that account's access |
| No recent version appears | The scheduled job didn't complete | Check job history and run a manual backup while the destination is reachable |
| Restore takes time before listing files | Remote metadata is being read without an existing local index | Let the documented recovery process finish; use the configured job when available |
| Restored file can't be opened | Wrong version, incomplete restore, or permissions | Read the summary, select the intended version, and restore to an accessible empty folder |
Variations for other backup goals #
Recover after losing the original installation
Use Restore and the documented direct-restore flow with destination details and the backup passphrase. That route reads enough remote information to build its file list. An exported configuration saves retyping settings. The local-database explanation describes why losing a local index doesn't automatically destroy the backup.
Back up to SFTP or cloud storage
Replace the local destination with the appropriate provider and its credentials. Use a dedicated directory or bucket path and check the destination before the first run. Storage-provider charges are separate. If you already use WinSCP, you can inspect an SFTP account independently while arranging its storage path.
Protect a whole Windows system
Use Veeam Agent for a system-image and recovery-media plan, or Rescuezilla for an offline disk image. Duplicati's folder workflow handles recoverable files and exports; it doesn't create the same bare-metal Windows recovery route.
Completion checklist #
- The selected source is the folder containing your current work.
- The destination is separate, reachable, and dedicated to this job.
- The initial backup completed and has a selectable version.
- You restored into another folder and opened recovered files.
- The passphrase, destination details, and exported configuration are available without the original computer.
- The next scheduled run can reach the destination.
Next steps #
Apply the same recovery-first thinking to your Thunderbird email and Aegis authenticator exports. If you want another snapshot and policy model, Kopia is an alternative to consider. Keep an ordinary manual archive separate from scheduled history; the 7-Zip vs PeaZip comparison covers that smaller task.
