10,000+ Photos on Your Hard Drive and Half Are Duplicates or Burst Shots? This Free Open-Source Rust Tool Reclaimed 76GB in 2 Minutes

If you have backed up family smartphones and work documents over the past five to ten years, your hard drive probably contains folders named Phone Backup 2021, Photos_Final_DoNotDelete, and New Folder (2)—consuming nearly 200 GB where almost half the files are duplicate copies or burst-shot near-duplicates.

Checking 40,000 photos manually is impossible, filename sorting fails whenever a copy was renamed (IMG_001.jpg vs Vacation (1).jpg), and commercial “duplicate cleaner” utilities lock the delete button behind a $49 paywall.

In our lab, we tested the 18,000-star open-source Rust utility Czkawka on a real 184.2 GB archive (42,600 photos, 4K videos, and PDFs)—reclaiming 76.4 GB (41.5% of the drive) in under two minutes.

1. Two Mathematical Scalpels: BLAKE3 Cryptographic Hashing + pHash Visual Similarity

1. Size Pre-Filter + BLAKE3 Cryptographic Hashing (For 100% Exact Clones): Instead of reading every byte on the disk, Czkawka groups files by exact byte size in 0.5 seconds, then computes ultra-fast multithreaded BLAKE3 hashes only on size collisions—finding 8,940 exact duplicate photos and 4K videos across 80 nested folders in 14.2 seconds regardless of filename.
2. Perceptual Hashing (pHash for Burst Shots & Resized Copies): In its Similar Images mode, Czkawka computes visual perceptual hashes to spot burst-mode shots and compressed thumbnails of the same scene, automatically highlighting lower-resolution copies so you always keep the highest-resolution original.

2. Foolproof 3-Step Cleanup (Zero Risk of Deleting Originals)

1. Add Target Folders: Add your photo/archive drives (D:\, E:\) in czkawka_gui.
2. Scan Duplicates & Similar Images: Run Duplicate Files (BLAKE3) first (reclaimed 54.8 GB in 14.2s), then Similar Images (reclaimed 13.4 GB in 28.5s).
3. Smart Select (Select all except biggest / oldest): Click Select → Select all except biggest, which guarantees that the highest-resolution original in every group remains untouched while sending only the redundant copies to the Recycle Bin.