Audiowmark Desktop

A PyQt6 desktop GUI for audiowmark - the robust audio watermarking tool by Stefan Westerfeld.

Embeds invisible, cryptographically secured watermarks into audio files. The watermark survives MP3/OGG re-encoding at 128 kbit/s or higher and can be detected from a copy without the original file.

Features

• Add Watermark - embed a watermark into WAV, MP3, or FLAC files with configurable strength
• Get Watermark - decode and verify watermarks; all keys in your key directory are tried automatically
• Key Management - generate and manage audiowmark keys and HMAC secrets
• Cryptographic payload - metadata (copyright, artist, title, purpose) is never stored directly in the audio; only its HMAC-SHA256 hash is embedded, making it one-way and unforgeable
• Database - maps payload hashes back to metadata for verified decoding
• MP3 input - MP3 files are decoded to a temporary WAV via ffmpeg before processing, working around libmpg123 frame-count limitations that would otherwise cause truncated output
• MP3 output and format conversion - output format is independent of input format; WAV, MP3, FLAC, and AIFF are all supported as output; MP3 output is re-encoded by ffmpeg at the original bitrate after watermarking
• Metadata preservation - audio tags (title, artist, album, genre, year, composer, comment, etc.) are copied from the input file to the output, across all supported formats (WAV RIFF INFO, MP3 ID3, FLAC Vorbis comments)
• Cover art preservation - embedded album art is carried through to the output without re-encoding
• Key rotation - multiple keys supported; all are tried automatically on decode

Security model

Property	This tool
Payload readable without key	No - private audiowmark key required
Payload forgeable	No - HMAC secret required
Watermark removable losslessly	No - private audiowmark key required
Metadata recoverable from payload alone	No - one-way HMAC
Multiple key rotation	Yes

Usage

1. Go to Key Management, choose a keys directory, and generate at least one audiowmark key and one HMAC secret.
2. In Add Watermark, select your input file, fill in the metadata fields (Copyright is mandatory), choose a key and secret, and click Add Watermark.
3. In Get Watermark, select a watermarked file - all keys are tried automatically and the matching metadata is displayed.

Watermark strength reference

Strength	Use case
10.0 (default)	Survives MP3/OGG at 128 kbit/s+ - good balance
13.0 – 15.0	Multiple conversions, low bitrate (64 kbit/s)
15.0 – 20.0	Maximum robustness; slight audibility possible

Installation

macOS - via .dmg (recommended)

Download the matching .dmg for your Mac from the Releases page, open it, and drag Audiowmark Desktop into the Applications folder.

File	Runs on
audiowmark-desktop_<version>_macos_arm64.dmg	Apple Silicon (M1 and later), native
audiowmark-desktop_<version>_macos_x86_64.dmg	Intel Macs, native; Apple Silicon via Rosetta 2

• Requires macOS 14 (Sonoma) or later
• audiowmark and ffmpeg are bundled - no separate installation needed

Note: The app is not notarized, as Apple requires a paid ($99/year) developer account for this. On first launch macOS may show "app is damaged". There are two ways to fix this:

Option 1 - Terminal (quickest): run once to clear the quarantine flag:

xattr -rd com.apple.quarantine "/Applications/Audiowmark Desktop.app"

Option 2 - System Settings: if macOS shows this dialog:

Click Done, then open System Settings - Privacy & Security and click Open Anyway:

Then confirm with Open Anyway once more.

Linux - via .deb package (recommended)

Download the latest .deb for your architecture from the Releases page:

sudo apt install ./audiowmark-desktop_<version>_amd64.deb
# or for ARM:
sudo apt install ./audiowmark-desktop_<version>_arm64.deb

The package bundles a compiled audiowmark binary - no separate installation needed.

Optional: install ffmpeg and ffprobe for MP3 input/output support and metadata/cover art preservation:

sudo apt install ffmpeg

Building from source

Requirements:

• Python 3.8+
• PyQt6 (pip install pyqt6)
• audiowmark in $PATH
• ffmpeg + ffprobe (optional, for MP3 input/output and metadata/cover art preservation)

git clone https://github.com/huddx01/audiowmark-desktop.git
cd audiowmark-desktop
pip install pyqt6
python3 src/audiowmark_gui.py

Building locally

macOS .dmg

bash packaging/build_mac.sh 1.0.0
# output: dist/audiowmark-desktop_1.0.0_macos_arm64.dmg   (on Apple Silicon)
#         dist/audiowmark-desktop_1.0.0_macos_x86_64.dmg  (on Intel)

Homebrew, Python, PyQt6, PyInstaller, and all native dependencies are installed automatically by the script.

Linux .deb

sudo apt-get install autoconf automake libtool pkg-config \
  libsndfile1-dev libmpg123-dev libzita-resampler-dev \
  libfftw3-dev libgcrypt20-dev \
  libavcodec-dev libavformat-dev libavutil-dev libswresample-dev \
  zstd fakeroot

bash packaging/build_deb.sh 1.0.0
# output: dist/audiowmark-desktop_1.0.0_amd64.deb

License

Copyright © 2026 Sojuzstudio, huddx01.

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

See LICENSE for the full text.

audiowmark itself is © Stefan Westerfeld, licensed under the LGPL v2.1+.