Skip to content

README.md

Latest commit

 

History

History
125 lines (98 loc) · 2.95 KB

README.md

File metadata and controls

125 lines (98 loc) · 2.95 KB

@wfloat/wfloat-web

@wfloat/wfloat-web is the browser package for Wfloat text-to-speech. Use it to turn text into spoken audio on your website.

Browser demo to hear how it sounds: https://wfloat.com/demo

Install

npm install @wfloat/wfloat-web
yarn add @wfloat/wfloat-web

Quick start

Your modelId is the Model Credential shown in your Wfloat account after sign up.

import { SpeechClient } from "@wfloat/wfloat-web";

const modelId = "your-model-credential";

await SpeechClient.loadModel(modelId, {
  onProgressCallback(event) {
    if (event.status === "downloading") {
      console.log("Downloading", Math.round(event.progress * 100) + "%");
      return;
    }

    if (event.status === "loading") {
      console.log("Initializing runtime");
      return;
    }

    console.log("Model ready");
  },
});

await SpeechClient.generate({
  text: "The signal is clean. Start the recording.",
  voiceId: "narrator_woman",
  emotion: "neutral",
  intensity: 0.5,
  speed: 1,
  silencePaddingSec: 0.1,
  onProgressCallback(event) {
    console.log("progress", event.progress);
    console.log("isPlaying", event.isPlaying);
    console.log("highlight", event.textHighlightStart, event.textHighlightEnd);
    console.log("chunkText", event.text);
  },
  onFinishedPlayingCallback() {
    console.log("Playback finished");
  },
});

API overview

  • SpeechClient.loadModel(modelId, { onProgressCallback }) loads the model onto the device. The first load downloads model and runtime assets for the browser.
  • SpeechClient.generate(options) generates a single utterance and starts playback.
  • SpeechClient.generateDialogue(options) generates multi-speaker dialogue from a list of segments.
  • SpeechClient.pause() and SpeechClient.play() control playback for the active request.

Progress callbacks

loadModel(...) emits:

{ status: "downloading", progress: number }
{ status: "loading" }
{ status: "completed" }

generate(...) emits:

{
  progress: number;
  isPlaying: boolean;
  textHighlightStart: number;
  textHighlightEnd: number;
  text: string;
}

generateDialogue(...) emits the same fields plus textHighlightSegment.

Dialogue example

await SpeechClient.generateDialogue({
  silenceBetweenSegmentsSec: 0.2,
  onProgressCallback(event) {
    console.log(event.progress);
  },
  onFinishedPlayingCallback() {
    console.log("Dialogue finished");
  },
  segments: [
    {
      text: "The door is locked.",
      voiceId: "narrator_man",
      emotion: "neutral",
    },
    {
      text: "Then we open it the loud way.",
      voiceId: "strong_hero_woman",
      emotion: "joy",
      intensity: 0.65,
    },
  ],
});

Browser note

Start generation from a user gesture such as a button click. Browsers can block audio playback until the page has received user interaction.

Contributing

Maintainer and local development notes live in CONTRIBUTING.md.