Configure Genesys TTS connector

Hi Rob, your synthesis endpoint is correct for a Speech resource in West Europe: https://westeurope.tts.speech.microsoft.com/cognitiveservices/v1. The voice-list endpoint is different and should be https://westeurope.tts.speech.microsoft.com/cognitiveservices/voices/list. Use the Speech resource key in the Ocp-Apim-Subscription-Key header and make sure the region in both URLs matches the resource that issued that key. For synthesis, Genesys also needs to send valid SSML with Content-Type application/ssml+xml and an appropriate X-Microsoft-OutputFormat value. Test both endpoints with curl or Postman first; if they work there but not in Genesys, export the connector's HTTP status/body and check whether Genesys is overriding the header names or request content type.

Hello @Rob van der Poel (RF) ,

Welcome to Microsoft Q&A .Thank you for reaching out to us.

Thank you for sharing the configuration details.

Based on the information provided, one Azure endpoint configuration issue has been identified. Azure AI Speech uses separate REST endpoints for speech synthesis and voice discovery, and the same endpoint cannot be used for both operations.

In addition, confirming the exact Genesys integration type will help ensure that the Azure Speech configuration aligns with the supported connector model and request/response requirements.

Configuration review

1. Azure Speech endpoint configuration

Synthesize Text URI

The current configuration is correct:

https://westeurope.tts.speech.microsoft.com/cognitiveservices/v1

This endpoint is used for Text-to-Speech synthesis requests and does not require any change.

List Voices URI

The current configuration is not correct for retrieving available voices.

Current value:

https://westeurope.tts.speech.microsoft.com/cognitiveservices/v1

Required value:

https://westeurope.tts.speech.microsoft.com/cognitiveservices/voices/list

Azure Speech provides separate REST operations for:

• Speech synthesis: https://westeurope.tts.speech.microsoft.com/cognitiveservices/v1

Voice discovery: https://westeurope.tts.speech.microsoft.com/cognitiveservices/voices/list

Using the synthesis endpoint for voice retrieval can prevent the connector from loading the available Azure Speech voice catalog successfully.

1. Authentication configuration validation The configured authentication method is supported. Required header: Ocp-Apim-Subscription-Key: Please verify the following:
   • The key is generated from the Azure AI Speech resource under Azure Portal > Speech Resource > Keys and Endpoint.
   • The key belongs to the same Speech resource being used for the Genesys integration.
   • The Speech resource region matches the endpoint region.
   • The key value does not contain additional spaces or formatting characters.
   • The value is an Azure AI Speech resource key and not an Azure subscription identifier.
   Example: Speech resource region: West Europe Endpoint: https://westeurope.tts.speech.microsoft.com
2. Validating Azure Speech endpoints independently Before further connector troubleshooting, validating the Azure Speech APIs directly helps confirm whether the Azure resource configuration and authentication are working correctly. A. Validate the voice list endpoint Test request:

      curl -X GET \
      "https://westeurope.tts.speech.microsoft.com/cognitiveservices/voices/list" \
      -H "Ocp-Apim-Subscription-Key:<Speech Resource Key>
   

   Expected result:
   • HTTP 200 response
   • JSON response containing available voice information
   Example:

      [
       {
       "Name": "en-US-JennyNeural",
       "Locale": "en-US",
       "Gender": "Female"
       }
      ]
   

   This confirms:
   • The endpoint is correct.
   • Authentication is successful.
   • The Speech resource is reachable.
   B. Validating speech synthesis Test request:

      curl -X POST \
      "https://westeurope.tts.speech.microsoft.com/cognitiveservices/v1" \
      -H "Ocp-Apim-Subscription-Key:<Speech Resource Key>" \
      -H "Content-Type: application/ssml+xml" \
      -H "X-Microsoft-OutputFormat: riff-24khz-16bit-mono-pcm" \
      --data "<speak version='1.0' xml:lang='en-US'><voice name='en-US-JennyNeural'>This is a test message.</voice></speak>" \
      --output test.wav
   

   Expected result:
   • HTTP 200 response
   • Generated audio file
3. Confirming Genesys integration type The next important validation point is confirming the exact Genesys integration currently being configured. The provided fields suggest that the integration may be using the Genesys TTS Connector or third-party TTS provider model. Genesys provides multiple TTS integration options, and the required configuration can vary depending on the selected integration. Please confirm whether the deployment uses:
   • Genesys TTS Connector
     • Third-party TTS Provider Integration
       • Dedicated Azure Cognitive Services Text-to-Speech Integration
   This confirmation is important because different integrations may have different requirements for:
   • Endpoint configuration
   • Authentication method
   • Request format
   • Audio response handling
   • Supported features such as SSML processing
4. Additional Azure validation checks If the connector continues to fail after updating the List Voices URI, please review the following areas. Azure Speech resource configuration: Verify:
   • The resource type is Azure AI Speech.
   • The resource region matches the configured endpoint.
   • The resource key belongs to the same Speech resource.
   Network configuration: Verify whether any of the following are enabled:
   • Firewall restrictions
   • Selected network access
   • Private Endpoint
   • Custom domain configuration
   Restricted networking settings may prevent Genesys Cloud from reaching the Azure Speech endpoint. Voice availability: After correcting the List Voices URI, confirm that the required voice exists in the returned voice list. The voice configured in Genesys should match the Azure Speech voice identifier. Example: en-US-JennyNeural

Reference for common responses:

• HTTP 200 → Request completed successfully.
• HTTP 400 → Invalid endpoint, parameter, or request format.
• HTTP 401 → Resource key, endpoint, or region mismatch.
• HTTP 403 → Access or authentication restriction.
• HTTP 415 → Unsupported content type.
• HTTP 429 → Quota or throttling limit reached.
• Timeout or connection failure → Network accessibility, firewall, DNS, or endpoint issue.

The following references might be helpful , please check them out

• Text to speech API reference (REST) - Speech service - Foundry Tools | Microsoft Learn
• How to synthesize speech from text - Speech service - Foundry Tools | Microsoft Learn
• Supported regions for Azure Speech - Foundry Tools | Microsoft Learn
• How to use private endpoints with Speech service - Foundry Tools | Microsoft Learn
• Language and Voice Support for Azure Speech - Foundry Tools | Microsoft Learn

Please let us know if the response was helpful

Thank you