Add Narration And Synchronize Actions

Add narration after the silent workflow is already correct.

1. Configure the narration provider and private service credential.
2. Add spoken text to a beat or a named take spanning beats.
3. Build and listen before scheduling actions.
4. Add anchors at meaningful spoken moments.
5. Join actions to those anchors with after.
6. Add an authored narration wait only when speech must wait for asynchronous work to complete.

The build reuses unchanged take audio. step=narration focuses the build on narration and its dependent presentation output while you revise speech.

See Narration, anchors, and waits for the exact syntax and Concepts for the timing model.