Orpheus TTS Podcast Generator

This project leverages Canopy AI's Orpheus TTS model to generate high-quality, multi-speaker podcasts. It integrates with Gemini to create engaging and dynamic conversational audio.

Features

• Multi-Speaker Support: Generate conversations between multiple characters with distinct voices.
• Gemini Integration: Utilizes Gemini to create podcast scripts and conversational content.
• Emotive Speech: Orpheus TTS supports emotive tags for realistic and engaging delivery.
• High-Quality Audio: Produces clear and natural-sounding audio at 24kHz.
• Multi-Language Support: Supports multiple languages (English, German, French, Spanish, Italian, Korean, Hindi, Chinese).

---

Demo: sebastianboehler.github.io/orpheus-podcast

License & Model Terms

Note: This project uses the Orpheus TTS models from Canopy AI. You must respect and comply with the license terms and usage restrictions provided by Canopy Labs for the Orpheus models. See the official release page for details.

Model Selection: Finetuned vs Pretrained

You can choose between finetuned and pretrained Orpheus TTS models for each language.

• Finetuned models are recommended for best quality and are the default.

To select the model type, use the --model-type argument or set the PODCAST_MODEL_TYPE environment variable.

Example:

python main.py --language german --model-type finetuned

or

export PODCAST_MODEL_TYPE=pretrained
python main.py --language french

Usage

1. Setup: Ensure you have the required dependencies installed (see requirements.txt).
2. Configuration: Configure the script with your desired speaker voices and Gemini prompts.
3. Generation: Run the script to generate your podcast audio.