A narration booth, driven by the timecode.

Performance direction, not just text-to-speech

Each beat carries a direction: pace, emphasis, pause length before a reveal, and the drop in energy that makes a reversal land. Those directions are compiled into the synthesis request, so the read has structure rather than uniform pleasantness.

• Held pauses before evidence reveals
• Emphasis marks on the beat's operative noun
• Breath and room tone preserved between paragraphs

Timing exported for the edit

Word-level timestamps come back with the audio. Cut points snap to phrase boundaries, captions are frame-accurate without transcription, and the motion stage knows exactly how long each shot must hold on screen.

Mix delivered to spec

The stem arrives loudness-normalised to platform target with de-essing and light compression applied, so the master needs no corrective pass before upload.