A lightweight client for running Orpheus TTS locally using OpenAI compatible API (e.g. those served via LM Studio, Llama.cpp/Ollama, or OpenWebUI).
- 🎧 High-quality Text-to-Speech using the Orpheus TTS model
- 💻 Completely local - no cloud API keys needed
- 🔊 Multiple voice options (tara, leah, jess, leo, dan, mia, zac, zoe)
- 💾 Save audio to WAV files
- Install LM Studio, Ollama, or Llama.cpp's Server
- Download the Orpheus TTS model (orpheus-3b-0.1-ft-q4_k_m.gguf) in your inference engine of choice
- For Ollama, do
ollama pull hf.co/isaiahbjork/orpheus-3b-0.1-ft-Q4_K_M-GGUF
- Start up the model For LM Studio
- Load the Orpheus model in LM Studio
- Start the local server in LM Studio (default: http://127.0.0.1:1234)
- Install dependencies:
python3 -m venv venv source venv/bin/activate pip install -r requirements.txt - (For OpenAI compatible endpoint users = Non-LM Studio) Set environment variables as needed
- export API_PATH=<your_server_url> (e.g. for Ollama, use http://localhost:11434/v1/chat/completions, for OpenWebUI, use http://your_server_url/api/chat/completions)
- export API_KEY=<your_api_key> (if you've set an API key to be required)
- export MODEL_NAME=<your_model_name> (if you've set a custom model name in your inference engine)
- Run the script:
python gguf_orpheus.py --text "Hello, this is a test" --voice tara
python gguf_orpheus.py --text "Your text here" --voice tara --output "output.wav"
--text: The text to convert to speech--voice: The voice to use (default: tara)--output: Output WAV file path (default: auto-generated filename)--list-voices: Show available voices--temperature: Temperature for generation (default: 0.6)--top_p: Top-p sampling parameter (default: 0.9)--repetition_penalty: Repetition penalty (default: 1.1)
- tara - Best overall voice for general use (default)
- leah
- jess
- leo
- dan
- mia
- zac
- zoe
You can add emotion to the speech by adding the following tags:
<giggle>
<laugh>
<chuckle>
<sigh>
<cough>
<sniffle>
<groan>
<yawn>
<gasp>Apache 2.0