fadupes is a CLI tool to help you find identical audio files in a music collection by scanning directories and files recursively and comparing audio characteristics.

• Recursive directory scanning
  • Accepts one or more directories and/or files as input
• Parallel processing
  • Uses multiple threads for faster scans
• Progress display
  • Global progress bar
  • Optional per-file ��currently scanning” output
• Resumable scans
  • Optional JSON state file
  • Automatically loads existing state
  • Periodically saved during processing
  • Saved automatically on Ctrl+C
• Symlink handling
  • Follows symlinks by default
  • Option to ignore symlinks
  • Avoids loop-back symlinks into scanned roots
• Filtering options
  • Ignore files by size (<, >, or range)
  • Skip files with a unique byte size for faster scans
• Logging
  • Duplicate groups written to identical_files.log
  • Processing errors written to identical_files_errors.log

You need Rust and Cargo installed.

cargo build --release

The binary will be available at:

target/release/fadupes

fadupes -i <path>
fadupes -i <directory1> <directory2>
fadupes -i <file1> <file2>

Inputs may be directories, files, or any combination of both.

• -i, --input <PATHS...>

  • One or more files or directories to scan

• --nolist

  • Disable per-file list output (keeps only the global progress bar)

• --nosym

  • Ignore symlinks instead of following them

• --checkpoint <N>

  • Save the resume JSON file every N scanned files
  • Default: 250

• -t, --threads <N>

  • Set the number of threads used for parallel scanning
  • Examples: -t5, -t 5, --threads 5

• --skip-unique-size

  • Skip files whose byte size appears only once (faster, but may miss duplicates)

• --ignore-size <EXPR>

  • Ignore files by size

  • Examples:

    • <3MB
    • >800MB
    • 3MB..800MB

• --state-file <PATH>

  • Path to the resume state file
  • Default: fadupes_state.json

• --no-resume

  • Disable loading and saving of the resume state

• State files are written in the current working directory by default
• If the state file exists, it is loaded automatically
• The state is saved periodically during the scan (tune with --checkpoint)
• On Ctrl+C, the state is saved before exiting

Each audio file is decoded and analyzed to extract audio properties. Files are considered identical if all of the following match:

• Total sample count
• Sample rate
• Bit depth
• Channel count
• Peak level
• RMS level (dB)

Files sharing the same characteristics are grouped together as duplicates.

• Console

  • Duplicate groups are printed to stdout

• Files

  • identical_files.log

    • Appended with duplicate file paths (grouped)

  • identical_files_errors.log

    • Created only if errors occur during processing

• Supported formats: WAV, FLAC
• Files larger than 800 MB are currently skipped

Run in debug mode:

cargo run -- -i /path/to/music

• Only WAV and FLAC are supported
• Duplicate detection is based on decoded audio characteristics, not tags
• No built-in deletion or interactive duplicate management

• Additional audio formats
• Interactive duplicate handling
• Persistent audio database of all scans