Talkdesk Conversation Orchestrator™ enables you to stream the agent and/or caller audio in real time. For more information, please visit our Conversation Orchestrator documentation.
- October 2, 2026 - Custom Request Headers on Bidirectional Voice Streams
- April 27, 2026 - Talkdesk Bring Your Own AI (SIP)
- September 28, 2022 - Talkdesk Conversation Orchestrator
October 2, 2026 - Custom Request Headers on Bidirectional Voice Streams
On October 2, 2026, we rolled out support for custom HTTP headers on the bidirectional WSS connection in the Connect to Autopilot Voice Studio component.
New:
-
Custom Request Headers: The Connect to Autopilot Voice component now supports a Headers field on its External Voice Stream connection. Define headers as a JSON object (for example, to pass an authentication token) or reference a single Studio variable instead. This replaces the previous workaround of encoding authentication parameters into the "Voice stream URL" itself.
Action required: None. This is optional and fully backward compatible. If the Headers field is left empty, the connection behaves exactly as it does today.
Resources:
For more information, please see this article.
April 27, 2026 - Talkdesk Bring Your Own AI (SIP)
On Monday, April 27th, 2026, Bring Your Own AI (SIP) became available for all customers as part of the Talkdesk Conversation Orchestrator.
New:
Talkdesk Bring Your Own AI (SIP): A new Talkdesk workflow that allows customers to integrate 3rd party voice AI bots using the Session Initiation Protocol (SIP), an open-standard VoIP protocol.
What’s included:
Bring Your Own AI (SIP) can be used to:
- Support bidirectional streams that can be used by other applications to do intent detection, speech-to-text, and voice biometrics, real-time voice and speech analytics, and other voice analysis or processing.
- Support a bidirectional stream where an Autopilot can participate in a conversation.
For additional information, please read this article.
September 28, 2022 - Talkdesk Conversation Orchestrator
On Wednesday, September 28th, 2022, we will release Conversation Orchestrator™.
New:
Conversation Orchestrator: This solution offers customers the possibility to stream the agent’s or the caller’s audio in real time.
What’s included
Conversation Orchestrator can be used to:
- Support multiple unidirectional streams that can be used by other applications to do intent detection, speech-to-text, and voice biometrics, real-time voice and speech analytics, and other voice analysis or processing.
- Support a bidirectional stream where an autopilot can participate in a conversation.
How to Access / Install
Conversation Orchestrator™ works by adding the “Stream audio unidirectional” component to a Studio flow to stream the audio of a given call to a web socket.