v0.1.0 — Initial release

🎉 First public release of VoxLoom — self-hosted voice AI studio

Run state-of-the-art text-to-speech and voice cloning on your own hardware.
No subscriptions. No rate limits. No per-character billing. No vendor lock-in.

What's in v0.1.0

• ✅ Offline system engine via pyttsx3 — zero setup, works everywhere, OS TTS quality
• ✅ kokoro interface scaffolded — Apache-2.0, EN + ZH, ~82M params
• ✅ scripts/make_demo.py — OmniVoice cloud-GPU demo generator
• ✅ Bundled demo/hello_system.wav — playable out-of-the-box
• ✅ MIT licensed, Apache-2.0 for bundled engines

Install

```bash
git clone https://github.com/huocan007/voxloom.git
cd voxloom
pip install -r requirements.txt
python voxloom.py tts "Hello world" --engine system --out hello.wav
```

Roadmap

• OmniVoice / Qwen3-TTS / Chatterbox cloud-GPU engines (scaffolded, need contributors)
• Web UI on top of the CLI
• Voice cloning from 3-second reference audio
• Optional hosted "VoxLoom Cloud" API

⭐ If this saves you from another rate-limit, star the repo. It helps.