UrBackup

Client-server backup for whole machines

Backup & Recovery AGPL-3.0 Intermediate ★ 932 stars

What is UrBackup?

UrBackup has a server that stores images and files, and clients that run on Windows, macOS and Linux. It handles image-level bare-metal restore as well as file backups, which is what makes it a fit for desktop machines.

Best for

Backing up and restoring whole desktop machines

Why choose UrBackup

UrBackup is built for a problem the file-oriented tools handle poorly: backing up whole desktop machines, including the ability to restore one to bare metal. It uses a client-server design, with a server holding the storage and clients running on Windows, macOS and Linux, and it can capture both image-level and file-level backups. The client is a small installer that non-technical users can run and then forget, which is exactly the right shape for a household or a small office with machines belonging to people who should not have to think about backups. Restoring a failed laptop from an image is a supported workflow rather than a manual reassembly of files, and the web interface lets you manage every machine from one place.

Replaces

  • Acronis
  • Veeam
  • Backblaze

Key features

  • Full image and file backups
  • Windows, macOS and Linux clients
  • Bare-metal restore from a bootable image
  • Web interface for administration and restore

What to watch out for

It is heavier than a single-binary tool: the server needs real storage, and image-level backups consume space quickly because they capture more than the files themselves. Running clients on every machine means installing and updating agents, and an agent that silently stops running produces a gap you will not notice until you need the backup. It assumes a persistent connection between clients and server, which means laptops that leave the network need their schedules and expectations configured deliberately. The interface is functional rather than polished, and its documentation assumes some familiarity with the concepts. Restoring an image to different hardware can surprise you if the drivers do not line up.

How to deploy

  • Server in Docker or installed directly
  • Install clients and approve them
  • Store backups on a dedicated volume

Getting started

Start the server with enough storage for at least one full image plus incrementals, and choose the storage location deliberately rather than accepting a default that may sit on the same disk. Install the client on one machine and run both a file backup and an image backup so you understand the difference in size and time. Configure retention so images do not accumulate indefinitely, since that is the most common way the storage fills. Turn on notifications for failed or skipped backups — a silent client is the failure mode that matters. Restore a file first, then practice a full restore on a spare machine or VM, because bare-metal recovery is not something to learn during an outage.

Typical setup

A server with storage on its own volume, sized for at least one full image plus incremental changes, and clients installed on each desktop machine that matters. File backups and image backups are both configured on the machine used to learn the system, so the difference in size and runtime is understood rather than guessed. Retention is set so images do not accumulate indefinitely, and notifications are enabled for failed or skipped runs because a silent client is the failure that goes unnoticed. A file restore has been performed, and a full image restore has been rehearsed on a spare machine or virtual machine.

Who should look elsewhere

Do not choose it for backing up a server, where a file-level tool with better deduplication is more appropriate and cheaper in storage. Avoid it if all you need is your documents directory, where it is a lot of machinery for a small job. And if you will not maintain agents on every machine, a backup system that assumes clients are running will quietly stop protecting exactly the machines you forgot about.

Project health

  • GitHub stars: 932
  • Last code push: 2026-09-29
  • Open issues: 19
  • Status: actively developed

Figures pulled from the GitHub API and refreshed periodically.

More in Backup & Recovery