Media (images, audio, video)

Text to Speech

Turn text into an audio file in a chosen voice with OpenAI, Azure AI Speech, Deepgram or ElevenLabs.

Text to Speech (media_tts) speaks text and stores the audio as a file.

Inputs

AI provider (integration_id): one of the workspace's integrations made from a media-capable connector — see Provider setup. It is chosen in the designer (not an expression). The picker lists only connections that can do the step, and warns when the chosen provider cannot do a feature you picked.

Text (text, required): expressions allowed.

Options

Setting	Key	Default	What it does
Voice	`voice`	the provider's default	A voice name or id from the provider (for example `alloy` for OpenAI)
Format	`format`	`mp3`	`mp3`, `wav`, `opus`, `aac` or `flac`
Speed	`speed`	1	The provider's range applies
Language	`language`	detect	Such as `en-US`
Style instructions	`instructions`	—	How to speak ("calm and friendly"), where the model supports it
Model	`model`	the provider's default
Output file name	`output_name`	`speech`
Output variable	`output_var`	`speech`

Outputs

Variable	Contents
`{{speech}}`	`file`, `characters`, `format`, `voice`, `provider`, `model`, `cost`
`{{speech_file}}`	The audio file

Example

A spoken notice

{
  "type": "media_tts",
  "label": "Speak notice",
  "config": {
    "integration_id": "<integration-id>",
    "text": "Hello {{customer.first_name}}, your order {{order.number}} has shipped.",
    "format": "mp3"
  }
}