Skip to content
dopemcp

Narration

The script writes itself.

Narration drafted from what is actually on screen, not from a generic template. Edit one line and regenerate one scene — the rest of the take stays put.

Grounded in the take

Written from the screen, not a template.

The agent watches what actually happened in the recording — which screens appeared, what got clicked, what the numbers said — and drafts narration to match. No generic "welcome to our product" filler, no lines describing a UI that isn't on screen.

  • Each line is tied to a scene, so audio never drifts from video
  • Edit one line and regenerate one scene — the rest stays put
  • Tone and length follow your prompt: launch-video punchy or docs-calm

Narration · demo-checkout

Draft 2
0:00

This is the new analytics dashboard — live numbers, no refresh.

0:07

Filter to the last 30 days and revenue updates in place.

0:12

The chart is worth a closer look.

editing…
0:18

Every row exports to CSV — one click, no dialog.

Change line three, regenerate scene three. Scenes one, two, and four don't move.

Then your voice reads it.

The finished script goes straight to text-to-speech through your own provider key — any voice in your library, in every language the project ships. See BYOK voice and localization.

Ready to make something dope?

Download the recorder, point your agent at it, and ship your first localized demo today.

Free during early access · No credit card

All platforms & requirements →