A lightweight client for running Orpheus TTS locally using LM Studio API.
- 🎧 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
- Download the Orpheus TTS model (orpheus-3b-0.1-ft-q4_k_m.gguf) in 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 - Run the script:
# Using a text string: python gguf_orpheus.py --text "Hello, this is a test" --voice tara # Or load text from a file: python gguf_orpheus.py --text-file path/to/your_text.txt --voice tara
python gguf_orpheus.py \
[--text "Your text here"] \
[--text-file "path/to/file.txt"] \
--voice tara \
--output "output.wav"
--text: The text to convert to speech--text-file: Path to a UTF-8 text file whose contents will be synthesized--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