A performance-capture stage, inside the editor.

Generation is not always the right answer. A performance carries timing and intent a prompt does not reach, and for the scenes that decide whether a game lands, recording one is often the shorter path to something that works.

This is a miniature performance-capture stage inside the Unreal editor. Microphone and webcam are recorded onto a single take clock, so voice and video stay in step, and every take is filed with its manifest rather than dropped into a folder to be identified later.

It runs two ways. Solo, take by take, when you are trying something out. Or as a planned session: lines queued from elsewhere turn the panel into a teleprompter so the performer reads rather than remembers, a producer plans the job, and a director reviews, plays back and chooses. Nothing downstream runs until a take has been picked — because recording and spending are not the same decision.

What makes it PlayableOps rather than a recorder is what happens next: the take is a candidate like any other. It flows into the same pipelines the generated content came from, carrying the record of which session and which choice produced it — so replacing it later is a decision rather than an excavation.

Deliberately narrow. Capture and take management only: face solving belongs to FaceForge, voices to SpeechForge, body motion to MotionForge. It records the performance; the other sets are what it drives.