Generate speech audio from text using Microsoft Edge's text-to-speech API.
Heavily inspired by rany2/edge-tts and SchneeHertz/node-edge-tts
- Using standard web APIs. Should work in all (modern) JS environment
- Provides subtitle/caption data
Using npm:
npm install @echristian/edge-ttsUsing pnpm:
pnpm install @echristian/edge-tts// Web
const { audio, subtitle } = await generate({
text: "Hello, world!",
voice: "en-US-JennyNeural",
language: "en-US",
});
// Create an audio element and play the generated audio
const audioElement = new Audio(URL.createObjectURL(audio));
audioElement.play();
// Access subtitle data
console.log(subtitle);Options that will be sent alongside the websocket request:
text(required): The text that will be generated as audiovoice(optional): Voice persona used to read the message. Defaults to"en-US-AvaNeural"- Please refer to Language and voice support for the Speech service
language(optional): Language of the message. Defaults to"en-US"- Please refer to Language and voice support for the Speech service
outputFormat(optional): Format of the audio output. Defaults to"audio-24khz-96kbitrate-mono-mp3"- Please refer to SpeechSynthesisOutputFormat Enum
rate(optional): Indicates the speaking rate of the text. Defaults to"default"- Please refer to Customize voice and sound with SSML
pitch(optional): Indicates the baseline pitch for the text. Defaults to"default"- Please refer to Customize voice and sound with SSML
Options for parsing the subtitle:
splitBy(required): The function will split the cues based on this option"sentence": will split the text usingIntl.Segmenter"word": will split the text to X count of words for each cue"duration": will split the text to X duration of milliseconds for each cue
count(optional): Used when splitting by"words"or"duration"- When splitting by
"words", count means the amount of words for each cue - When splitting by
"duration", count means the duration in milliseconds for each cue
- When splitting by
metadata(required): Array of metadata received throughout the websocket connection