Using the Text-to-speech API

This quickstart guide offer ready-to-use cURL commands for the Text-to-speech API.

Overview

Using the Text-to-speech (TTS) API, you can generate an audio output from a text prompt or text file input. Options with this API allow you to:

1. Use a voice of your choice from a catalog of stock voices.
2. Choose from a wide variety of voices across genders, age groups, and speaking styles.

Prerequisites

Review the Getting Started page for this API for authentication and setup.

API credentials

You'll need:

• client_id
• client_secret

Quickstart

In the cURL command below, be sure to update:

• Authorization with the bearer token.
• x-api-key as per the prerequisite.
• mediaType as per input type.
• url with the generated pre-signed URL.
• voiceId specifies the unique ID of the voice to be used for speech generation. Users should refer to the Voices List API to choose the appropriate voice ID.

The command returns a response object like the one below. Use the statusUrl from the response to check the job result.

{
    "jobId": "986fc222-1118-4242-b326-eb9873e3982f",
    "statusUrl": "https://audio-video-api.adobe.io/v1/status/{jobID}"
}

Generate a video from plain text prompt

curl --location 'https://audio-video-api.adobe.io/v1/generate-speech' \
-H 'Authorization: Bearer <Token>' \
-H 'x-api-key: <Client_ID>' \
-H 'Content-Type: application/json' \
--data-raw '{
    "script": {
        "text": "<script text>",
        "mediaType": "text/plain",
        "localeCode": "en-US"
    },
    "voiceId": "<voice ID>",
    "output": {
        "mediaType": "audio/wav"
    }
}'

Generate using a text file

curl --location 'https://audio-video-api.adobe.io/v1/generate-speech' \
-H 'Authorization: Bearer <Token>' \
-H 'x-api-key: <Client_ID>' \
-H 'Content-Type: application/json' \
--data-raw '{
    "script": {
        "source": {
            "url": "<pre-signed URL of text file>"
        },
        "mediaType": "text/plain",
        "localeCode": "en-US"
    },
    "voiceId": "<voice ID>",
    "output": {
        "mediaType": "audio/wav"
    }
}'

Check the status of a job

Use the GET Result API to see the status of a job. In the command below, update:

• statusUrl
• Authorization
• x-api-key

The statusUrl is returned in the response of the TTS API call.

curl --location '<statusUrl>' \
  -H 'Authorization: Bearer <Token>' \
  -H 'x-api-key: <Client_ID>' 

Sample TTS API response

{
    "jobId": "986fc222-1118-4242-b326-eb9873e3982f",
    "status": "succeeded",
    "output": {
        "destination": {
            "url": "<pre-signed URL of the result>"
        }
    }
}

The url can be used to download the generated audio for the input.

Verify with Content Credentials

Adobe participates in the content authentication initiative for AI-generated assets, addressing concerns around content legitimacy. Register your content by uploading the file at ContentCredential.org.