Skip to main content
Not comfortable on camera? Don’t have time to film? Prompt to Video generates a realistic talking-head video from a script or voiceover, using your AI Twin or a pre-built actor. You never have to be on screen.

​
How Prompt to Video works

You provide a script or voiceover audio. Captions generates a complete talking-head video using either your AI Twin (a digital avatar that looks and sounds like you) or one of Captions’ built-in actors. A realistic person delivers your content, synced to your words. The output is a finished video with captions, B-roll, music, and motion graphics applied. If you want the avatar to look like you, set up your AI Twin before starting. Record a 1-minute calibration video and Captions generates your digital avatar. See the AI Twin setup guide for exact instructions.
You can skip this and use one of Captions’ built-in actors instead. Your AI Twin just makes the output look like you personally.

​
Create a no-camera video with Prompt to Video

1

Open Prompt to Video

From the Captions home screen, tap Prompt to Video.
2

Write your script or record a voiceover

Type your script directly into the script field, or record an audio voiceover to use as the driver. You can also import an existing audio file. The script or audio becomes the spoken content of the video.
Scripts of 60-90 seconds produce the best results. Very short scripts (under 20 seconds) or very long ones (over 2 minutes) may not generate as well.
3

Choose your actor

Select My AI Twin if you’ve set one up, or choose from Captions’ library of built-in actors. Preview each actor to find the right fit for your content’s tone.
4

Choose an AI Edit style

Pick a style that controls the look of the final video: caption design, B-roll, music, transitions, and motion graphics. Browse the 95+ style options and select one that matches your content. You can use the same style across all your Mirage videos for visual consistency.
For educational or informational content, choose a clean, minimal style. For high-energy content, choose a dynamic style with more motion graphics.
5

Generate your video

Tap Generate. Captions produces the complete talking-head video synced to your script or audio. Generation typically takes a couple of minutes depending on video length.
6

Review and adjust

Watch the output from start to finish. Use Co-editor to make quick changes by typing: “fix the caption on the word X,” “swap the B-roll at 0:15,” or “change the music.” You can also edit manually in the timeline.
Co-editor is the fastest way to make small corrections without re-generating the whole video.
7

Export

Tap Export when you’re happy with the result. The video exports at 1080p in vertical 9:16 format, ready for TikTok, Reels, or YouTube Shorts.

​
Tips

  • Record your voiceover in a quiet space even though you’re not on camera. Clean source audio produces a better-synced, more natural result
  • 60-90 seconds is the sweet spot for Prompt to Video; longer scripts can be split into separate videos
  • Use the same actor and style across your videos to build a consistent visual identity
  • If you’re not happy with the first output, try re-generating. Slight variations in timing and expression are normal

​
What’s next?

AI Twin Setup

Create a digital avatar that looks and sounds like you.

Prompt to Video docs

Full reference for the Prompt to Video feature.

AI Edit Workflow

The full AI Edit workflow for creators.

Co-editor

Chat to edit. Type what you want changed.
Last modified on April 20, 2026