Create an AI Twin of yourself from a recording, then generate new videos of it reading any script you write.
Plan: Pro, Grow, Webinar, and Business
Device: Computer Browser
This feature uses AI Credits
Please note:
This feature is in beta and is being released gradually to Riverside users.
Step by step
- Log in to your Riverside account.
- Navigate to the relevant studio.
- From the side bar on the left, click AI Twins.
-
Under Start from a recording, choose one of your recordings
Or, click Upload to pick a video file from your computer.Choose a recording that meets these requirements:
- You're the only person on camera.
- Your face is clearly visible and stays in frame for the entire recording.
- Nothing covers your mouth, including your hands, a microphone, or other objects.
- Your voice is clearly audible, with at least 8 seconds of speech.
- In the Name field, type a name for your avatar.
- Click Create AI Twin.
- Review the consent notice, then click Agree and create.
Training your AI twin can take up to 3 minutes. - Your avatar appears on the Your avatars page with a short preview.
- On your avatar's card, click Generate video.
- In the script field, type or paste what your avatar will say.
Scripts can be up to 5,000 characters long. - At the bottom left corner of the script field, click the current aspect ratio to choose an orientation for your video.
- Click Generate video.
Rendering can take up to 11 minutes, depending on your script's length.
You'll get an email when your video is ready.
Good to know
- Accounts on the Pro, Grow, and Webinar plans can create 5 AI Twins per studio. Business accounts can create unlimited AI Twins.
- Once your video is ready, find it in Projects, under a project named after your avatar.
- Uploaded files must be in MP4 or MOV format.
- Magic Clips isn't available for videos generated with an AI Twin. Magic Episode can be generated.
- AI twins can currently only be created from your own recordings or your own uploaded audio files. Creating an AI twin of someone else, including from a guest's async recording, is not supported.