terran-adjutant-tts

StarCraft 2 Terran Adjutant voice synthesis powered by onnxruntime-node + espeak-ng.

Requirements

• Node.js >= 16

Install

npm install terran-adjutant-tts

Usage

const PiperTTS = require('terran-adjutant-tts')

const tts = new PiperTTS({ outputDir: './output' })

const result = await tts.processStreamWithAudio(
  "Additional supply depots required.",
  './output/speech.wav'
)

API

new PiperTTS(config?)

| Option | Default | Description | |--------|---------|-------------| | modelPath | bundled model | ONNX model path | | modelConfigPath | bundled config | Model config path | | outputDir | ./output | Output directory | | chunkSize | 500 | Token chunk size |

processStreamWithAudio(text, outputPath?, config?)

Synthesizes text to WAV. Uses onnxruntime-node for inference + espeak-ng (WASM) for phonemization.

Docker

docker build . -t tts-native
docker run --rm tts-native node -e "
  var P = require('./index.js');
  var p = new P({ outputDir: '/tmp' });
  p.processStreamWithAudio('Hello.', '/tmp/out.wav')
    .then(r => console.log('OK', r.modelPath))
"

Model

• Voice: StarCraft 2 Terran Adjutant (custom trained)
• Format: ONNX VITS (Piper)
• Sample Rate: 22050 Hz

License

MIT