How to use openai.audio.speech.create in node.js

API
1

The specs say that the return value is “The audio file content.” and examples show “mp3.arrayBuffer()” and “mp3.stream_to_file()”.

Can the documentation be updated to explain exactly what is returned from the speech.create() mthod?

1 Like
2

Isn’t the docs just implying it returns the raw bytes the content of an audio file?

3

Looks like this one worked.

const mp3 = Buffer.from(await response.arrayBuffer());

So the client.audio.speech.create method seems to return a Node “response” object, and we get the raw bytes using the .arrayBuffer() method.

4

Docs need to be updated, please, in Nodejs almost all elements on the response interface are undefined.

PLEASE OPENAI, update audio docs :frowning:

5

Check the reference page

  const mp3 = await openai.audio.speech.create({
    model: "tts-1",
    voice: "nova",
    input: "Today is a wonderful day to build something people love!",
    speed: 1.0,
  });

  const buffer = Buffer.from(await mp3.arrayBuffer());
  await fs.promises.writeFile(filename, buffer);

Ensure that there is “await” in that last line. That is a promise.