Levante Audio

Levente Audio is a graphical user interface for the ElevenLabs API. It can also utilize OpenAI's Whisper for speech-to-text transcription, if installed.

Features

With Levante Audio, you can:

• Interact with the ElevenLabs API in real-time
• Record or upload audio files for transcription via OpenAI's Whisper (optional)
• Convert text to audio using ElevenLabs

Installation

Ensure you have Python 3.9 or higher installed. Creating a Python virtual environment before the installation is recommended.

To install the application, first clone the repository:

git clone https://github.com/digital-pro/levante-audio.git

Then navigate into the directory and install the package using pip:

cd levante-audio
pip install .

To use OpenAI's Whisper API or a local Whisper implementation for transcription, you can install the extras like this:

pip install .[whisper_api]

Or:

pip install .[whisper_local]

Or if you want both:

pip install .[whisper_api,whisper_local]

🚧 Important notes for Linux and macOS users 🚧

Tkinter Installation

For Linux: Make sure tkinter is installed for your Python environment. You can do this by installing the python3-tk package using your package manager. For example, if you're using Ubuntu, you can install it with:

sudo apt-get install python3-tk

For macOS: Install tkinter via Homebrew:

brew install [email protected]

OpenSSL Issue (macOS only)

If you encounter an error with urllib3: You may need to install or update OpenSSL. Use Homebrew to install OpenSSL:

brew install openssl

If you've already installed OpenSSL but you're still encountering the error, reinstall Python linked with the Homebrew version of OpenSSL:

brew reinstall python

If the error persists, try installing an older version of urllib3:

pip3 install 'urllib3<2.0'

Configuration

To use ElevenLabs you need to get an API Key. You can get one for free, but they do limit how many characters you can translate before you need to pay.

To use this application, you need to set the environment variable: ELEVENLABS_API_KEY to the API key you've gotten from ElevenLabs. Levante Audio will read it. Obviously we don't want the keys in the code, as the code is public.

Usage

Run the main.py script to start the application:

python main.py

You'll see a long table of our current translations in the middle column of the screen.

First: Choose "Synthesize Voices" and you'll see voice control options below, and the Voice selection drop-down will become active.

Second: Select a voice from the drop-down. NOTE: I think there are other voices, but they don't show up when the code asks for a list of voices.

Third: Click on the translation you'd like to preview (from the scrolling table)

When you do that, the translated text should appear in the text box.

Fourth: If you want to edit the translation, add a pause (using an embedded tag like ), or change the clarity & stability sliders, do that now.

Fifth: To hear how the translation is vocalized, press "Generate Speech".

License

This project is licensed under the terms of the MIT license. Serious kudos to winedarkmoon who wrote the underlying code.

​