GUIDE

Text to speech in Power Automate: an HTTP action and an MP3

Power Automate flows already move text around: emails, Forms answers, SharePoint items, Teams messages. To turn some of that text into speech, the flow needs one call to a text to speech API and a step that stores the audio.

This guide shows how to do it with the HTTP action and the SuperClonica speech API.

Get a free API key Pick a voice Or run the engine yourself

Before you start

The HTTP action is a premium connector in Power Automate, so the flow needs a plan that includes premium connectors. You also need a SuperClonica API key: sign in with Google or GitHub and create it in the Speech API section of your workspace; the free plan gives a minute of audio a day.

Add the HTTP action

After the trigger, add an HTTP action and fill it in:

  • Method POST, URI https://superclonica.com/api/v1/speech.
  • Headers: Authorization = Bearer and your key, Content-Type = application/json. Better still, keep the key in an environment variable or Azure Key Vault.
  • Body: {"text": "[dynamic content with the text]", "voice": "M1", "language": "en", "format": "mp3"}.

Store the audio

The answer’s body is the MP3 itself. Add OneDrive or SharePoint “Create file”, name it with the .mp3 extension and use the HTTP action’s Body as the file content. From there the flow can attach it to an email, post it to Teams or link it in a list item. Add an Idempotency-Key header, for example the item’s ID, so a retried run never voices the same text twice.

Errors, limits and privacy

Configure the HTTP action’s retry policy or a scope with “run after” for failures: 429 means the plan’s 10-minute window is used up and 503 that the queue is full; both answers carry Retry-After. A request takes up to 20,000 characters. The text is processed on SuperClonica servers, so for confidential documents use the SuperClonica engine on your own machine instead.