Speech-to-text (STT) engines overview

AI-generated summary

Genesys Cloud provides comprehensive speech-to-text (STT) capabilities through multiple implementation approaches designed to support voice bot conversations with real-time and offline transcription. The platform offers three primary options: included STT engines, bring your own technology (BYOT) integrations, and Bot Transcription Connector for third-party automatic speech recognition services via the AudioHook protocol. Included STT engines available within Architect bot flows comprise six industry-standard providers: Google Cloud Speech-to-Text, Microsoft Azure Cognitive Services STT, Amazon Transcribe, Genesys Native STT, Deepgram Flux, and Deepgram Nova-3. These engines are incorporated into bot pricing and support mixed resource billing with Agentic Virtual Agents. Advanced features including custom ASR dictionary support and advanced ASR configuration are available exclusively through Amazon Transcribe and Genesys Native engines. Organizations with data residency or compliance requirements should review regional endpoint documentation, as Deepgram engines may process audio outside the customer's Genesys Cloud region. BYOT STT integrations enable organizations to incorporate supported Microsoft Azure, Google Cloud, or legacy Nuance services into Architect bot flows, with pricing structured on per-turn or per-minute rates. These integrations facilitate voice-enabling external chatbots through Dialog Engine Bot Flows, where transcribed customer utterances are transmitted via data actions to external systems, with text-to-speech converting responses back to audio. Configuration of STT solutions involves installing integrations from Genesys AppFoundry, which adds applications to the organization's Admin > Integrations page for subsequent activation and configuration. Bot Transcription Connector streams audio from Dialog Engine Bot Flows to third-party ASR engines and returns transcripts to the bot flow, allowing selection of third-party engines as default options or for specific slot collection operations. Genesys does not charge for Bot Transcription Connector usage; however, customers are responsible for all third-party STT vendor fees, processing costs, and development expenses. Standard bot flow usage charges for Dialog Engine Bot Flows continue to apply under regular platform pricing, with AWS regions supported for Google Cloud TTS and STT integrations.

Speech-to-text (STT) engines recognize spoken audio and transcribe it into text in real time or offline. Genesys Cloud supports multiple STT options for voice bot conversations, including Genesys-provided engines, bring your own technology (BYOT) integrations, and Bot Transcription Connector.

You can use an STT integration to voice-enable an external chatbot. In a Genesys Dialog Engine Bot Flow, the STT engine transcribes the customer’s utterances. Data actions can send the transcribed text to the external chatbot, and text-to-speech converts the chatbot response to audio for the customer.

STT options for Genesys Cloud

Genesys includes STT engines from multiple speech recognition providers. You can also use BYOT STT engines, which enable you to bring your own STT or ASR engines and then integrate them in Architect bot flows. Finally, use the Bot Transcription Connector to stream audio from bot flows to a third-party automatic speech recognition (ASR) engine and receive transcripts back using the Genesys AudioHook protocol.

Option	Use this option when
Included STT engine	You want to use an STT engine available directly through Genesys Cloud.
BYOT STT engine	You want to use your own supported Microsoft Azure, Google Cloud, or legacy Nuance service.
Bot Transcription Connector	You want to connect Genesys Cloud to a supported third-party transcription service through AudioHook.

Included STT engines

Genesys Cloud includes industry-standard STT engines. These speech-to-text engines are only available in Architect bot flows. Genesys includes these engines in Bot pricing, and as mixed resource billing with Agentic Virtual Agents that consume the same interaction as the bot flow. The following table lists each STT and its service provider:

STT engine name	Vendor name	Custom ASR dictionary support	Advanced ASR configuration
Google Cloud	Google Cloud Speech-to-Text
Microsoft Azure	Microsoft Azure Cognitive Services STT
Amazon Transcribe	AWS Transcribe Speech-to-Text
Genesys Native	Genesys native STT
Deepgram Flux	Deepgram Flux Speech-to-Text
Deepgram Nova-3	Deepgram Nova-3 Speech-to-Text

Notes:

When you use Deepgram Flux or Deepgram Nova-3, Deepgram can process audio outside your Genesys Cloud region. Genesys recommends that customers with data residency or compliance requirements review the Deepgram regional endpoint documentation before they enable either engine in production.
To use a custom ASR dictionary, you must choose Amazon Transcribe or Genesys Native.
Genesys named the STT engines to more accurately reflect their underlying speech recognition providers. For more information, see Addition of Deepgram Flux and Nova-3 speech-to-text (STT) engines and updated STT engine names.

BYOT STT engines

Genesys enables you to bring your own STT or ASR engines and then integrate them with Genesys Cloud into Architect bot flows. For more information about BYOT STT engines pricing, see Bring your own technology services model (per turn/minute rates). The following BYOT STT integrations are available:

Microsoft Azure Cognitive Services Speech-to-Text (STT) integrationClick to expand

Install the Microsoft Azure Cognitive Services integration from Genesys AppFoundry. Installation adds the app to your organization’s Admin > Integrations page. Then configure and activate the integration.

Google Cloud Speech-to-Text (STT) integrationClick to expand

Bot Transcription Connector

Bot Transcription Connector streams audio from Genesys Dialog Engine Bot Flows to a third-party ASR engine and returns the resulting transcripts to the bot flow. You can select the third-party ASR engine as the default engine or for specific slot collection using an Ask for Slot action. These tasks explain how to install, configure, and activate Bot Transcription Connector from Genesys Cloud’s Integrations page.