Text-to-speech (TTS) engines overview

AI-generated summary

Genesys Cloud provides comprehensive text-to-speech capabilities through multiple engine options, including a native Genesys TTS engine, Enhanced TTS engine available from the Genesys AppFoundry at no additional cost for Dialog Engine Bot Flows, and third-party integrations with Amazon Polly, Google Cloud, Microsoft Azure, ElevenLabs, and Nuance Text-to-Speech. The platform supports generative TTS voices across call flows, bot flows (virtual agents, AI guides, and agentic virtual agents), and voice surveys, with available voice models including Google Cloud's Chirp 3: HD voices and Gemini 3.1 Flash TTS, ElevenLabs' Eleven Flash v2.5, and Amazon Polly's generative voices. Organizations can configure multiple TTS engines simultaneously and independently select specific voices and languages across different flows, enabling scenarios where a flow uses a different TTS voice than its default language setting—such as using English (en-US) for data playback with an English (en-IE) voice for Irish accent output. Flow authors can configure TTS engines and voice options per language within individual flows through Architect, with the capability to select voices supported by the configured engine. Technical implementation requires Edge and Media Tier version 1.0.0.7920 or later. Regional and security restrictions apply: only third-party TTS solutions are supported in the US East 2 (Ohio)/FedRAMP region, and secure call flows support only PCI-certified solutions including Genesys TTS engine, Genesys Enhanced TTS, Amazon Polly, Google Cloud Text-to-Speech, Microsoft Azure Cognitive Services Text-to-Speech, and Nuance Text-to-Speech. When third-party TTS engines are deactivated, the system automatically defaults to the Genesys TTS engine for supported languages; if unsupported, text-to-speech playback will not occur. Key considerations include ensuring TTS voices match the flow's language, removing audio recordings from prompts when implementing TTS voices, and recognizing that playback performance varies by language, dialect, and voice provider. The platform supports both standard TTS integration and a Bring Your Own Technology (BYOT) model, with migration pathways available for transitioning between providers. Additional resources cover TTS engine fundamentals, Architect workflow tools, prompt management, and detailed migration guidance for transitioning between TTS solutions.

Feature coming soon: Genesys Generative TTS

PrerequisitesClick to expand

Note: The organization's default text-to-speech (TTS) engine is Genesys TTS. To access additional third-party TTS engines in Architect, first obtain them from the Genesys AppFoundry. Then, configure them in Genesys Cloud. For more information, see .

To meet your organization’s text-to-speech (TTS) needs, Genesys offers the following options:

Genesys provides an enhanced TTS engine that you can install from the Genesys AppFoundry. The Genesys Enhanced TTS engine offers many voice and language options and great audio quality. Genesys includes usage of the Enhanced TTS engine free of charge in Dialog Engine Bot Flows.
You can add third-party TTS engine integrations and then select voice and language options. These integrations expand language options and enable you to select a TTS voice for the organization, serving callers across built-in applications with the most appropriate voice. Architect can integrate with these third-party solutions for text-to-speech playback on a per-flow basis.
Genesys Cloud also includes a default Genesys text-to-speech (TTS) engine that offers more limited voice and language options and is suitable for testing purposes.

For a specific flow, you can configure the TTS engine and voice options for each language you include in the flow. You can select any TTS voice that the configured TTS engine supports. For more information, see Select a TTS engine and voice for a flow.

Genesys Generative TTS

Genesys Generative TTS converts text into human-like voices across supported languages. Architect flow authors can work with generative TTS voices for natural sounding text-to-speech playback within call flows, bot flows, and voice surveys.

Genesys supports the following voice models under Genesys Generative TTS:

Voice model	Provider	Call flows	Bot flows (Virtual Agents, AI guides, and agentic virtual agents)	Voice surveys
Chirp 3: HD voices	Google Cloud	✓	✓	✓
Eleven Flash v2.5	ElevenLabs	✓	✓	✓
Gemini 3.1 Flash TTS	Google Cloud	✓	✓	✓
Generative voices	Amazon Polly	✓	✓	✓

Notes:

TTS voices that do not directly match the flow’s language can result in issues with the customers’ text-to-speech experience.
If your flow uses prompts, Architect plays the prompt’s audio recording based on the flow’s language. If you want to use a TTS voice for prompts, remove the audio recording and add text instead.
Because not all TTS engines operate the same way, Genesys and third-party TTS engine playback performance can vary depending on language, dialect, and voice. Perform testing to ensure that you find the best solution for your use case, or contact your solutions consultant. For more information about third-party TTS engine performance, see Test your third-party TTS engine playback.
Only third-party TTS solutions are supported in the US East 2 (Ohio)/FedRAMP region.
Only PCI-certified third-party solutions are available in Architect secure call flows. Secure call flows can only use the Genesys TTS engine, Genesys Enhanced TTS, Amazon Polly TTS, Google Cloud Text-to-Speech, Microsoft Azure Cognitive Services Text-to-Speech, or Nuance Text-to-Speech.
You can configure an Architect flow with a language that supports runtime data playback and select a different TTS voice for localized speech. For example, you can use English (“en-US”) as the flow language to enable runtime data playback and select English (“en-IE”) voice to provide an Irish accent.

Warning: Use caution when deactivating a third-party TTS engine integration. Currently, an administrator can deactivate a third-party TTS engine in Genesys Cloud without a dependency check when the TTS engine is the default engine for the organization or selected in a flow.

If an administrator deactivates a third-party TTS engine chosen as the default TTS engine for the organization or any flows, the system defaults to the Genesys TTS engine for supported languages. If Genesys TTS does not support the language, then the text-to-speech string does not play at call flow runtime. For more information, see Search for flows that use a TTS engine.

[NEXT] Was this article helpful?

Get user feedback about articles.

Genesys Generative TTS

Genesys supports the following voice models under Genesys Generative TTS:

Voice model	Provider	Call flows	Bot flows (Virtual Agents, AI guides, and agentic virtual agents)	Voice surveys
Chirp 3: HD voices	Google Cloud	✓	✓	✓
Eleven Flash v2.5	ElevenLabs	✓	✓	✓
Gemini 3.1 Flash TTS	Google Cloud	✓	✓	✓
Generative voices	Amazon Polly	✓	✓	✓

Notes:

TTS voices that do not directly match the flow’s language can result in issues with the customers’ text-to-speech experience.
If your flow uses prompts, Architect plays the prompt’s audio recording based on the flow’s language. If you want to use a TTS voice for prompts, remove the audio recording and add text instead.
Because not all TTS engines operate the same way, Genesys and third-party TTS engine playback performance can vary depending on language, dialect, and voice. Perform testing to ensure that you find the best solution for your use case, or contact your solutions consultant. For more information about third-party TTS engine performance, see Test your third-party TTS engine playback.
Only third-party TTS solutions are supported in the US East 2 (Ohio)/FedRAMP region.
Only PCI-certified third-party solutions are available in Architect secure call flows. Secure call flows can only use the Genesys TTS engine, Genesys Enhanced TTS, Amazon Polly TTS, Google Cloud Text-to-Speech, Microsoft Azure Cognitive Services Text-to-Speech, or Nuance Text-to-Speech.
You can configure an Architect flow with a language that supports runtime data playback and select a different TTS voice for localized speech. For example, you can use English (“en-US”) as the flow language to enable runtime data playback and select English (“en-IE”) voice to provide an Irish accent.

Text-to-speech (TTS) engines overview

Genesys Generative TTS

[NEXT] Was this article helpful?

Loading...

Text-to-speech (TTS) engines overview

Genesys Generative TTS

[NEXT] Was this article helpful?