TransparentMeta

TransparentMeta is the standard for compliance with AI audio transparency legislation with metadata.

It's an open-source Python library developed by Transparent Audio. It allows users to add and verify cryptographically signed metadata in audio files.

TransparentMeta helps generative AI companies comply with AI transparency laws, such as the EU AI Act and the California AI Transparency Act. By using cryptographic signing, robust file validation, and a modular SDK, users can label AI-generated audio content with transparent metadata simply and effectively.

🗣️ Community

• 🗨️ Join the Transparent Audio Discord for support and discussion
• 🐛 If you discover bugs, please report them in the GitHub issue tracker

🚀 Features

• Add cryptographically signed transparency metadata to MP3 and WAV files
• Read and verify metadata to detect AI-generated audio
• Key pair generation and signature verification built-in
• Simple Python SDK for easy integration in your application
• 100% test coverage via unit and integration tests

🔧 Use Cases

• Speech synthesis and voice cloning labeling output with transparency metadata
• Generative AI music and sound effects tagging output with metadata and attribution
• Help with compliance with EU AI Act & California AI Transparency Act
• Building trusted pipelines with verifiable audio provenance

📦 Installation

From PyPi

Installing TransparentMeta from PyPI using pip is the recommended way:

pip install transparentmeta

From Source

You can also install TransparentMeta directly from the repository. This project uses Poetry for dependency management. You’ll need to install Poetry first.

To install TransparentMeta from source code run:

git clone https://github.com/Transparent-Audio/transparentmeta.git
cd transparentmeta

Then, activate your Poetry environment:

poetry env use python3.12  # or any Python >=3.12 you have installed
poetry env activate

The last command will print something like:

source /path/to/virtualenv/bin/activate

Copy and paste that command into your shell to activate the Poetry virtual environment.

This is the workflow for Poetry ≥2.0. If you are using Poetry ≤1.x, run poetry shell instead — it will activate the environment in a new subshell automatically.

Now, you can install the project with:

make install

To install in development mode, run:

make install_dev

make install and make install_dev wrap relevant Poetry commands. They can be found along with other useful commands in the Makefile.

📚 Getting Started

• Check the getting started guide in the documentation to pick up the basic functionality of TransparentMeta in 5 minutes
• Check the /examples folder for quick usage examples. Start by reading the README.md file inside the folder
• For a more thorough introduction to TransparentMeta, check the tutorials in the documentation
• If you're a visual learner, watch the video tutorials on YouTube

📖 Documentation

Check full documentation on Read the Docs.

📂 Project Structure

• transparentmeta/ - Main library code
• examples/ - Example scripts and usage
• tests/ - Unit and integration tests
• docs/ - Sphinx documentation
• Makefile - Useful commands for development

🧩 Dependencies

TransparentMeta relies on the following core libraries:

• mutagen – for reading and writing audio metadata (MP3, WAV, etc.)
• cryptography – for generating and verifying digital signatures
• pydantic – for data validation and data structures

You can find all dependencies in pyproject.toml.

✅ Running tests and quality checks

To run tests, linting, and type checks, use:

make checklist

🐍 Python version

TransparentMeta supports Python 3.12 and above. Ensure you have a compatible version installed.

📝 License

This project is licensed under GPL-3.0 or later. See the LICENSE file for details.

✍️ Contribution

While we don’t currently accept contributions, we plan to set up the necessary infrastructure to do so in the future. If you’re interested in contributing, please contact us.

📬 Contact

If you have any pressing questions or issues, please write to Transparent Audio's CTO Valerio Velardo at valerio@transparentaudio.ai.

The best way to get support or request features is by joining the Transparent Audio Discord