Orpheus TTS Node SDK

Node.js SDK for streaming speech from Orpheus TTS.

Install

npm install @canopylabsai/orpheus-tts

Quick Usage

import { OrpheusClient } from "@canopylabsai/orpheus-tts";

const client = new OrpheusClient({
  apiKey: process.env.ORPHEUS_TTS_API_KEY,
  client: "demo",
  provider: "baseten",
});

await client.connect({ voice: "leobaldo", websocketCount: 1 });

for await (const chunk of client.stream("Hello world", { voice: "leobaldo" })) {
  // chunk is PCM audio data (Buffer)
}

await client.close();

Example Scripts

Single run:

node examples/test_orpheus_connect_by_voice.js \
  -v "leobaldo" \
  --provider baseten \
  --client demo \
  --api-key YOUR_API_KEY

Parallel (16 by default):

node examples/test_orpheus_connect_by_voice_parallel.js \
  -v "leobaldo" \
  --provider baseten \
  --client demo \
  --api-key YOUR_API_KEY

Publishing to npm

From node-package/, use the publish script:

bash scripts/publish-npm.sh

Or via npm script:

npm run publish:package

Dry run (package validation only, no publish):

DRY_RUN=1 bash scripts/publish-npm.sh

Make sure npm auth is configured first (npm login or NPM_TOKEN flow).