How I Built an Incremental iPhone Backup Tool for Windows Using Tauri and Rust
If you shoot high-bitrate 4K videos on an iPhone and edit them on a Windows PC, you already know the pain: transferring gigabytes of footage via Windows File Explorer over MTP is unreliable. Connections randomly drop mid-transfer, transfers freeze, and resuming requires manually checking which files already made it to your external drive.
I wanted a lightweight, single-purpose, local-first utility without iTunes bloat or cloud subscriptions. So I built iphone-backup-sync, an open-source desktop application powered by Tauri v2, Rust, and Vue 3.
The Architecture
Instead of relying on Windows' built-in MTP layer, the Rust backend interfaces directly with the device over USB using libimobiledevice and ifuse.
The frontend is built with Vue 3 and TypeScript, keeping the UI minimal and responsive while Rust handles the heavy lifting of file system scanning and byte-level comparisons.
Key Engineering Decisions
- Incremental Syncing: Before transferring anything, the backend scans the target directory on your Windows drive and compares it against the iPhone's DCIM structure. Already-backed-up files are skipped automatically.
- Non-Destructive Conflict Handling: Interrupted transfers often leave behind partial or corrupted files. If the app finds an existing file with the same name but a different byte size, it never overwrites your existing backup. Instead, it saves the incoming file with a
_newsuffix so you can inspect both safely. - Automated CI/CD Releases: Every version tag (
v*) triggers a GitHub Actions workflow that compiles and attaches clean Windows.exeinstallers directly to the GitHub Releases page.
Try It Out or Contribute
The project is 100% open-source. You can inspect the Rust/Tauri source code or grab the latest Windows binary directly from GitHub:
🔗 Repository: github.com/mahdis24/iphone-backup-sync
I am currently working on automatic USB plug-in detection and expanding support for Android devices. If you run into Windows driver edge cases or have ideas for improvements, feel free to open an issue or drop a comment below!