piper-plus-g2p

Multilingual G2P (Grapheme-to-Phoneme) for TTS. eSpeak-ng free. MIT licensed. 8 languages.

Why piper-plus-g2p?

MIT licensed -- no eSpeak-ng (GPL) dependency in your TTS pipeline
8 languages -- JA, EN, ZH, KO, ES, FR, PT, SV with consistent IPA output
IPA-first design -- returns pure IPA token sequences; encoding to model-specific phoneme IDs is a separate step

Quick Start

Add to your Cargo.toml:

[dependencies]
piper-plus-g2p = { version = "0.1", features = ["naist-jdic"] }

use piper_plus_g2p::{Phonemizer, PhonemizerRegistry};
use piper_plus_g2p::english::EnglishPhonemizer;

let mut registry = PhonemizerRegistry::new();
registry.register("en", Box::new(EnglishPhonemizer::new().unwrap()));

let phonemizer = registry.get("en").unwrap();
let (tokens, prosody) = phonemizer
    .phonemize_with_prosody("Hello, world!")
    .unwrap();

// Encode tokens to phoneme IDs for a Piper ONNX model:
// let ids = piper_plus_g2p::encode::tokens_to_ids(&tokens, &phoneme_id_map)?;

Feature Flags

Flag	Default	Description
`english`	on	Enable English phonemizer
`chinese`	on	Enable Chinese phonemizer
`korean`	on	Enable Korean phonemizer
`spanish`	on	Enable Spanish phonemizer
`french`	on	Enable French phonemizer
`portuguese`	on	Enable Portuguese phonemizer
`japanese`	off	Enable Japanese phonemizer (pulls in `jpreprocess`)
`naist-jdic`	off	Bundle the NAIST-JDIC dictionary for Japanese (implies `japanese`)
`all-languages`	off	Enable all language backends including `japanese`

To use only specific languages, disable defaults:

[dependencies]
piper-plus-g2p = { version = "0.1", default-features = false, features = ["english", "japanese"] }

Supported Languages

Language	Code	Feature flag	Backend
Japanese	`ja`	`japanese`	jpreprocess (NAIST-JDIC)
English	`en`	`english`	Rule-based (CMUdict-derived)
Chinese	`zh`	`chinese`	Pinyin-to-IPA
Korean	`ko`	`korean`	Rule-based
Spanish	`es`	`spanish`	Rule-based
French	`fr`	`french`	Rule-based
Portuguese	`pt`	`portuguese`	Rule-based
Swedish	`sv`	`swedish`	Rule-based

Piper Model Compatibility

Use PiperEncoder to convert IPA tokens to phoneme IDs for Piper ONNX models:

use piper_plus_g2p::encode::{PiperEncoder, UnknownTokenMode};

// Load phoneme_id_map from model's config.json
let encoder = PiperEncoder::new(phoneme_id_map, UnknownTokenMode::Strict)?;
let phoneme_ids = encoder.encode(&tokens)?;

C FFI (Mobile Bindings)

Enable the ffi feature for C-compatible functions suitable for iOS (Swift) and Android (Kotlin) bindings via UniFFI:

piper-plus-g2p = { version = "0.1", features = ["ffi", "english"] }

Cross-Platform Consistency

Also available as:

Python: piper-plus-g2p on PyPI
npm: @piper-plus/g2p for browser/WASM
Go: go get github.com/ayutaz/piper-plus/src/go/phonemize

All implementations share the same PUA mapping and are validated against a common test fixture.

Minimum Supported Rust Version

1.88

License

MIT

piper-plus-g2p 0.2.0