Examples

Scan Multiple Directories

Scan several directories together — duplicates across directories are detected:

dublette 2020 2021 2022 --dry-run

Use shell glob expansion to match directories by pattern:

dublette 202* --dry-run
dublette ~/Photos/*/RAW --dry-run

This is especially useful for finding photos that were copied across year folders or backup directories.

Clean Up a Photo Library

Preview duplicates in your photos directory:

dublette ~/Photos --dry-run

Review the output, then delete:

dublette ~/Photos --yes

To also remove any 0-byte corrupted files:

dublette ~/Photos --delete-empty --yes

Deduplicate a Video Archive

Process only video files:

dublette ~/Videos --only videos --dry-run

Videos are compared by extracting a frame at the 1-second mark and hashing it. This catches re-encoded copies and different containers of the same content.

Images Only (Skip Video Processing)

If you do not have ffmpeg installed, or want to avoid the overhead of frame extraction:

dublette ~/Media --only images

Audio: Keep the Best Copy of Each Recording

The default recording match groups the same recording across formats and bitrates, keeping the highest-fidelity copy (a flac survives its mp3 rip):

dublette ~/Music --only audio --dry-run

To only remove exact copies that differ in tags – never grouping different encodings:

dublette ~/Music --only audio --audio-match encoding --dry-run

Adjust the Threshold

The default threshold of 1 is conservative. To require exact perceptual matches:

dublette ~/Photos --threshold 0 --dry-run

To catch more aggressively similar files (e.g., same photo with different crops or slight edits):

dublette ~/Photos --threshold 3 --dry-run

Always preview with --dry-run when increasing the threshold.

Scripting with JSON Output

Use --json and --dry-run to get machine-readable output:

dublette ~/Photos --dry-run --json > results.json

The exit code indicates whether duplicates were found:

dublette ~/Photos --dry-run --json --quiet
if [ $? -eq 1 ]; then
  echo "Duplicates found"
fi

Parse the JSON output with jq:

dublette ~/Photos --dry-run --json | jq '.groups[].duplicates[]'

Count total duplicates:

dublette ~/Photos --dry-run --json | jq '.total_duplicates'

CI / Automated Pipeline

Combine --json, --quiet, and --yes for non-interactive use:

dublette /data/media --json --quiet --yes > /tmp/dedup-report.json

Use the exit code in a CI step to detect duplicates without deleting:

dublette /data/uploads --dry-run --quiet
EXIT_CODE=$?
if [ "$EXIT_CODE" -eq 1 ]; then
  echo "::warning::Duplicate media files detected"
fi

Verbose Output for Debugging

To see the computed hash for each file and the hamming distance for every comparison:

dublette ~/Photos --verbose --dry-run 2> hashes.log

Verbose output is written to stderr, so it can be captured separately from the table or JSON output on stdout.

Suppress Progress Bars

When piping output or running in a non-interactive terminal:

dublette ~/Photos --quiet --dry-run

This suppresses the progress bars and scanning status messages while still printing the results table.

Combine Flags

Delete all duplicates and empty files without prompts, with JSON output, no progress bars:

dublette ~/Photos --delete-empty --yes --json --quiet

Preview image-only duplicates with strict matching and verbose logging:

dublette ~/Photos --only images --threshold 0 --dry-run --verbose