Photocopier

Back up your photos and videos to external drives — organised by date, de-duplicated, and verified.

Photocopier copies your media onto a backup drive into a tidy TYPE/YYYY-MM/ structure, never copies the same photo twice (even across runs and messy source folders), verifies every copy byte-for-byte before trusting it, and can prove to you that everything made it before you delete the originals.

dest/
├── JPG/
│   ├── 2024-05/2024-05-14_18-32-01_IMG_2041.jpg
│   └── 2024-06/…
├── HEIC/2024-06/…
├── CR3/2024-06/…
└── MP4/2024-06/…

Install

Requires Python 3.10+ and exiftool (for photo dates):

brew install exiftool pipx     # macOS
pipx install photocopier-cli   # the command is still `photocopier`

Usage

1. See what you have

photocopier stock ~/Pictures

Reports how many files photocopier recognises as media, and lists any extensions it would skip (check those for anything you care about, e.g. .zip archives of photos).

2. Back up

photocopier copy ~/Pictures /Volumes/MyBackupDrive/photos

• Files are organised by the date the photo was taken (EXIF), falling back to the file's modification date.
• Duplicate content is skipped — a record of everything ever copied lives in copied_hashes.txt inside the destination. Don't delete that file.
• Every copy is verified: the destination bytes are re-read and compared to the source before the file is recorded as done.
• Interrupted runs (unplugged drive, Ctrl-C) are safe: in-flight files are cleaned up and re-copied next time.
• Add --dry-run to preview what would be copied without writing anything, and -v to list each file as it's handled.

Run it again any time — already-copied files are skipped, so incremental backups are fast and safe.

3. Verify before deleting anything

photocopier verify ~/Pictures /Volumes/MyBackupDrive/photos

Exit code 0 means every media file in the source is present in the destination — it is safe to delete the source. Any missing or unreadable files are listed explicitly.

For a paranoid check that re-reads every file on the destination drive (catching files that were corrupted after they were copied):

photocopier verify --deep ~/Pictures /Volumes/MyBackupDrive/photos

Auditing two backup drives against each other (recommended if you keep two copies):

photocopier verify --deep /Volumes/DriveA/photos /Volumes/DriveB/photos
photocopier verify --deep /Volumes/DriveB/photos /Volumes/DriveA/photos

Both exiting 0 means the two archives hold identical content.

Supported formats

Photos: jpg, jpeg, png, gif, heic, heif, webp, avif, tif, tiff, bmp RAW: cr2, cr3, dng, nef, arw, orf, raf, rw2 Video: mp4, mov, m4v, mkv, avi, mpg, mpeg, 3gp, 3g2, mts, m2ts, vob, flv, wmv, webm

How de-duplication works

Each file's content is hashed (MD5, read in chunks). Hashes of everything ever copied are stored in copied_hashes.txt in the destination. A file whose content hash is already recorded is skipped, regardless of its name or folder — so re-organising your source folders never causes duplicate copies.

The hash algorithm is MD5 for continuity with existing archives; it's used for de-duplication, not security.

Development

git clone https://github.com/dsmurrell/photocopier
cd photocopier
python -m venv .venv && ./.venv/bin/pip install -e ".[dev]"
./.venv/bin/pytest

Disclaimer

Photocopier is free, open-source software provided as is, without warranty of any kind (see LICENSE). It is designed so you never have to take it on trust — every copy is verified byte-for-byte, and the verify command lets you prove a backup is complete before deleting anything. Please follow the golden rules anyway: keep two backup copies on separate drives, and only delete originals after photocopier verify reports ✓.

License

MIT