piper-phonemize

Try it Online

You can try piper-phonemize directly in your browser using WebAssembly:

• Hugging Face Spaces: https://huggingface.co/spaces/csukuangfj/piper-phonemize
• ModelScope Studios (for users in China, 中文用户): https://modelscope.cn/studios/csukuangfj/piper-phonemize/summary

Introduction

Phonemization library for Piper text-to-speech system.

Converts text to IPA phonemes using espeak-ng.

Supports Python, Go, Rust, Swift, Pascal, JavaScript (Node.js/WASM), C/C++, Tauri, Android, and iOS.

Installation

Python

pip install piper_phonemize -f https://k2-fsa.github.io/icefall/piper_phonemize.html

Go

go get github.com/csukuangfj/piper-phonemize-go/piper_phonemize

Rust

Add to your Cargo.toml:

[dependencies]
piper-phonemize = "1.4.8"

The crate automatically downloads prebuilt libraries and embeds espeak-ng-data. No manual setup required.

Swift

Add to your Package.swift:

dependencies: [
    .package(url: "https://github.com/csukuangfj/piper-phonemize", branch: "master")
]

JavaScript / Node.js

WASM version (works everywhere, no native dependencies):

npm install piper-phonemize

Native addon (better performance, requires platform-specific binary, which is installed automagically):

npm install piper-phonemize-node

C / C++

Download pre-built libraries from GitHub Releases or build from source.

Quick Start

Python

import piper_phonemize

# Initialize espeak-ng
piper_phonemize.initialize("/path/to/espeak-ng-data")

# Phonemize text
result = piper_phonemize.phonemize("Hello world", "en-us")
print(result)

Go

package main

import (
    "fmt"
    pp "github.com/csukuangfj/piper-phonemize-go/piper_phonemize"
)

func main() {
    pp.Initialize("") // uses embedded espeak-ng-data
    result := pp.Phonemize("Hello world", "en-us")
    defer pp.DeletePhonemizeResult(result)
    for i := 0; i < result.GetNumSentences(); i++ {
        phonemes := result.GetPhonemes(i)
        fmt.Printf("Sentence %d: %v\n", i, phonemes)
    }
}

Rust

use piper_phonemize::phonemize_to_string;

fn main() {
    // No initialization needed - espeak-ng-data is embedded
    let sentences = phonemize_to_string("Hello world", "en-us").unwrap();
    for s in &sentences {
        println!("{}", s);
    }
}

JavaScript / Node.js

WASM version:

const piperPhonemize = require('piper-phonemize');

// espeak-ng-data is bundled — no initialization needed!
const result = piperPhonemize.phonemizeToString('Hello world', 'en-us');
console.log(result); // ['həlˈoʊ wˈɜːld']

Native addon:

const piperPhonemize = require('piper-phonemize-node');

// espeak-ng-data is bundled — no initialization needed!
const result = piperPhonemize.phonemizeToString('Hello world', 'en-us');
console.log(result); // ['həlˈoʊ wˈɜːld']

C

#include "c-api.h"

int main() {
    PiperPhonemizeInitialize("/path/to/espeak-ng-data");

    PiperPhonemizeResult *result = PiperPhonemizeText("Hello world", "en-us");
    if (result) {
        int32_t num_sentences = PiperPhonemizeResultGetNumSentences(result);
        // Process phonemes...
        PiperPhonemizeDestroyResult(result);
    }
    return 0;
}

Tauri Example (Cross-Platform GUI)

A cross-platform desktop/mobile app built with Tauri v2. See tauri-examples/ for details.

Pre-built apps are available on the GitHub Releases page:

Building from Source

Using CMake

cmake -Bbuild -DBUILD_SHARED_LIBS=ON
cmake --build build --config Release

Using Docker

docker buildx build . -t piper-phonemize --output 'type=local,dest=dist'

License

GPL-3.0-or-later (due to espeak-ng dependency)

Links

• Piper TTS
• espeak-ng
• sherpa-onnx
• GitHub Issues