# Prompts for the demo

The demo is one twin saying one line on HeyGen's three engines, with the mouths beside the real footage. These prompts make your own.

## Make yours

> Use $create-ai-avatar to make my digital twin from the eye-contact takes of my last video. Say the price first. Then render "Hey, it's me. This is my digital twin, trained on my own takes, and speaking in my own ElevenLabs voice. Pretty wild, right?" in my voice on Avatar V, IV and III, and show me the mouths beside my real footage.

## Fix a twin that over-acts

> My avatar's mouth is too expressive. Render the same line, with the same audio and no motion prompt, on Avatar IV and Avatar III, and put the mouths beside my real takes.

## A talking photo instead

> Use $create-ai-avatar to make this photo of me say this line in my ElevenLabs voice on Avatar IV. No twin: just the photo and the audio.

## Remix the page

> Make a page like the skill's demo for my own twin: a pick for each engine, the clip with a sound button, the mouth sheet with the picked row lit, and the price of each.

The line in the demo is 8.8 s, in Meng Voice 2026 Pro on ElevenLabs v4. The twin is "Meng To · 31 Oct takes", trained from the eye-contact takes of his Framer video.