#6767th of 67 in VoiceFor voice, speech and music
orpheus-tts
Llama-3B-based text-to-speech model with emotion tags and streaming output
- Stars
- 6.4k
- License
- Apache-2.0
- Last commit
- Dec 2025
- Language
- Python
5,980 on 4 Apr 20266,385 on 31 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Orpheus TTS is a speech synthesis model built on a Llama-3B backbone, released as pretrained and finetuned English checkpoints plus a multilingual research family on Hugging Face. The orpheus-speech Python package streams audio through vLLM, with a stated latency of about 200ms. The repo also includes finetuning and pretraining scripts, a watermarking option, and a llama.cpp path for running without a GPU.
Who it is for: Developers building realtime TTS apps or finetuning custom voices
Strengths
- Streaming generation with stated ~200ms latency, ~100ms with input streaming
- Emotive tags such as <laugh>, <sigh> and <gasp>, plus 8 named English voices
- Finetuning scripts and dataset prep notebook; README cites good results from ~50 examples
- Apache-2.0 code with a llama.cpp option for no-GPU inference
Weaknesses
- Streaming package occasionally skips frames, listed as an open checklist item
- Voice cloning Colab notebook is listed as needing a fix
- vLLM version problems; README suggests pinning vllm==0.7.3 as a workaround
- Smaller 1b, 400m and 150m models unreleased; multilingual models are research preview
What it needs
- GPU optional
- Needs vllm, transformers, trl, accelerate, wandb (finetuning)
- Models: orpheus-tts-0.1-finetune-prod, orpheus-tts-0.1-pretrained, Orpheus multilingual research models
Quick answers
Is orpheus-tts open source?
Yes. Its code is on GitHub under the Apache-2.0 license.
Does orpheus-tts need a GPU?
No. It runs on a CPU and can use a GPU when you have one.
Also in Voice
See all 67| Rank | Project | Score |
|---|---|---|
| 1 | voicestudioDesktop app for voice cloning, dubbing, dictation and audiobooks | 97 out of 100 |
| 2 | whisper.cppC/C++ port of OpenAI Whisper for local speech-to-text transcription | 89 out of 100 |
| 3 | handyOffline push-to-talk dictation app that pastes transcripts into any text field | 89 out of 100 |
| #4 | xiaozhi-esp32ESP32 firmware for a voice chatbot with MCP device and cloud tool control | 82 out of 100 |
| #5 | buzzOffline desktop app for transcribing and translating audio with Whisper | 81 out of 100 |