Speech Cloud Documentation

What is IVONA Speech Cloud?

Welcome to the IVONA Speech Cloud Developer’s Guide.

IVONA Speech Cloud offers an easy way to add speech to your application. Using simple HTTP requests, you specify a text string and a voice, and IVONA Speech Cloud returns your text as spoken by the specified voice.

Service Highlights

IVONA Speech Cloud features include:
  • Simple model - A single authenticated request is sufficient to produce an audio stream. A ready-to-use Java client library is available for download.

  • Low latency - IVONA Speech Cloud is designed to minimize processing time. The following design points reduce latency:

    • Minimal communication round trips - Low communication overhead reduces overall latency.

    • Streamable response - The first usable part of the response is sent as soon as possible, before all speech request processing has finished.

    • Geo-distributed service endpoints - Your requests can be handled by regional endpoints distributed geographically.

  • Security - IVONA Speech Cloud offers proven security through state-of-the-art cryptographic methods to authenticate users, prevent unauthorized usage, and encrypt communication.

  • Scalability and Reliability - Service is constantly monitored for any anomalies. Most outages can be prevented and the rest fixed with minimal downtime.

  • Customizable pronunciations - The API fully supports the Pronunciation Lexicon Specification (PLS), enabling you to quickly adjust the pronunciation to your needs.

  • Beneficial pricing model - Monthly billing for the number of units used.

    • The number of units is calculated by dividing the number of characters in the input text by 200 and rounding up the result. SSML tags are removed from the calculation.

    • The total number of units requested per account is limited to 50,000 monthly. If the limit is reached, the account is disabled until the end of the current month and automatically enabled at the beginning of the next month. If you want to increase the limit, please contact the IVONA Sales Team.


To minimize service response latency, an encoded audio data stream is sent in chunks, with each chunk sent to service clients as soon as it is available, not after the entire response is assembled. IVONA Speech Cloud implements this using a standard HTTP/1.1 mechanism called Chunked Transfer Encoding.

Example service response:

HTTP/1.1 200 OK
x-amzn-RequestId: <RequestId>
Content-Type: audio/mpeg
Transfer-Encoding: chunked
Date: <Date>


Each data chunk may vary in size and is preceded with a size value.

Supported Operations

IVONA Speech Cloud offers the following set of operations:


The CreateSpeech operation performs a synthesis of input text to an audio stream containing the spoken text. Detailed information can be found in the IVONA Speech Cloud API Reference


The ListVoices operation returns a list of TTS voices available for speech synthesis using the CreateSpeech activity. Detailed information can be found in the IVONA Speech Cloud API Reference.

PutLexicon, GetLexicon, DeleteLexicon and ListLexicons

These operations enable you to manage user-defined lexicons available for speech synthesis. Detailed information can be found in the IVONA Speech Cloud API Reference.

List of IVONA Speech Cloud endpoints


The list of endpoints is formatted as follows:

aws region: URI (physical location)

To leverage the mentioned functionalities, you need to connect to one of the provided endpoints:

  • eu-west-1: tts.eu-west-1.ivonacloud.com (EU, Dublin)

  • us-east-1: tts.us-east-1.ivonacloud.com (US East, N. Virginia)

  • us-west-2: tts.us-west-2.ivonacloud.com (US West, Oregon)

IVONA Speech Cloud credentials

To use IVONA Speech Cloud, an Access Key and Secret Key pair is required. You can generate keys by logging into your IVONA account and visiting the Credentials page.


IVONA Speech Cloud is not part of AWS. This means that credentials retrieved from AWS IAM do not work with our service.

Copyright © 2015 IVONA Software. All rights reserved. Terms of Use | Privacy Policy