Duration of the generated video as reported by the Agents API for Talks (V2) and Clips (V3)
agents. 0 when the call produced no discrete video — on Expressive (V4) agents, and in a
text-only chat mode.
OptionalsessionId of the session this call was made on; the SDK sends it back on later streaming requests.
What the server said about the request. Not a fixed set — the Agents API declares it as an
open string — so treat an unrecognized value as accepted; a rejected request comes back as
an HttpError. 'success' on the stub the SDK returns where no video is produced.
Id of the generated video.
Use it to correlate this call with the video the agent then plays, which onVideoStateChange reports the start and end of, and which interrupt() cancels. Empty when the call produced no discrete video — on Expressive (V4) agents, and in a text-only chat mode.
What speak() resolves with: the video the agent is about to stream.
The fields come from the Talks (V2) and Clips (V3) API, which generates a discrete video while answering the request. Where no such video exists the call still resolves, with the same stub —
status'success',duration0and an emptyvideoId: on Expressive (V4) agents, whose speech is streamed over the data channel rather than rendered as a separate video, and in a text-only chat mode (ChatMode.TextOnly, ChatMode.Playground or ChatMode.Maintenance), which produces no video at all.