Metadata-Version: 2.3
Name: zipvoice-rs
Version: 0.1.0
Summary: Python ctypes bindings for zipvoice-rs
Author: thewh1teagle
Author-email: thewh1teagle <61390950+thewh1teagle@users.noreply.github.com>
Requires-Python: >=3.12
Description-Content-Type: text/markdown

# zipvoice-rs

Python bindings for ZipVoice GGUF inference.

The published wheels bundle the native `zipvoice-capi` dynamic library for the
supported platforms, with the native backend optimized per platform:

- macOS Apple Silicon: Metal + Accelerate
- Linux x86_64: Vulkan
- Linux aarch64: Vulkan
- Windows x86_64: Vulkan

## Install

```bash
uv pip install zipvoice-rs
```

## Use

```python
from zipvoice_rs import ZipVoice

with ZipVoice("zipvoice-en-q8_0.gguf", "vocos-mel-24khz-q8_0.gguf") as model:
    model.generate_wav(
        "prompt.wav",
        "ɹˈiəl tʃˈeɪndʒ bɪɡˈɪnz wˈɛn jɔːɹ hˈoʊp bɪkˈʌmz stɹˈɔŋɡɚ ðæn jɔːɹ ɛkskjˈuːsᵻz.",
        "ðə mˈɔːɹnɪŋ tɹˈeɪn ɚˈaɪvd bɪsˈaɪd ði ˈoʊld stˈoʊn bɹˈɪdʒ.",
        "out.wav",
    )
```

See `examples/` for English, Hebrew, and mixed Hebrew/English phonemization
examples.

For local development and native-library packaging notes, see `BUILDING.md`.
