Make a Photo Talk with AI Avatars
Turn a headshot into a talking video. Pick a photo or make your own avatar, add speech from your History or new lines in any voice, then style captions, add zooms, and export up to 4K.
Avatars turns a photo into a talking video. Pick a headshot and some speech, and you get back a video of that person saying it, with their lips moving in time with the words. Use your own photo, pick one of the public avatars, or make your own avatar from a few choices and a short description.
Each video can be up to 20 seconds long. You can make it at 480p, 720p or 1080p, then style the captions, add a hook and zooms, and export. Avatars runs on the desktop app and in your browser with VCP Cloud.
What the desktop app needs
On the desktop app, Avatars needs an NVIDIA graphics card with 8 GB of VRAM and 16 GB of RAM, with the models stored on an SSD. The first video downloads the model, and the Review step shows how much it downloads and how much free disk space it needs. If your computer can't run Avatars, you can use it in your browser with VCP Cloud instead.
How It Works
Open Avatars in the sidebar. Setup has four steps across the top: Photo, Speech, Subtitles and Review. Your projects stay in the Avatars list, so you can come back to any of them.
1. Pick a Face
Start with one of these:
- Your own photo - Drop a photo or browse for one. JPG, PNG and WebP work.
- Generate - Make a new person from a place, a framing and a description. See Create an Avatar.
- Public avatars - Pick a ready-made person and one of their looks.
- A saved avatar - Pick any avatar you made or used before.
The best photos show one person facing the camera, shoulders up, in good light and without sunglasses. If the photo has more than one person in it, you are asked to crop it down to one face.
2. Crop the Photo
When you upload a photo, it is shown on the Photo step with Replace photo and Crop buttons. Click Crop to pick a shape (Vertical, Square or Landscape), drag the photo to move it, and zoom with the slider, the mouse wheel or a pinch. The video takes the shape of your photo, so crop to vertical for TikTok, Reels and Shorts, or to landscape for YouTube.
3. Add the Speech
On the Speech step, choose where the words come from:
- From library - Pick a clip from your History. Search the list and play a clip before you pick it. Clips over the length limit are marked.
- New speech - Type what they should say and voice it here. Choose TTS to use one of your voices, or Design to describe a new voice, then pick the Voice model. Some models also take a style prompt. The box shows about how long the speech will run.
To voice lines with a cloned voice, clone it first, then pick it under TTS.
4. Style the Subtitles
Pick a caption style on the Subtitles step and see it on your photo as you go. Pick None to make the video without captions. You can change the style later in the editor. The styles and settings are the same as in the Subtitle Generator.
5. Review and Generate
The Review step lists your speech and subtitle choices, each with an Edit button to jump back. Then:
- How they move - Optionally describe how the person acts while they talk, such as "Smiling calmly and nodding now and then".
- Resolution - Choose 480p, 720p or 1080p. 1080p is offered when your computer or plan can make it.
- Consent - For your own photos, tick I have the right to use this person's likeness. Avatars you generate or pick from the public avatars skip this.
Click Generate video. The screen shows each stage and about how long is left, and you can cancel at any time. You can leave the screen while it works. The project shows Generating in your list until it is Ready.
On VCP Cloud, the Review step shows the estimated cost in tokens before you start.
What You Can Do
When the video is ready, it opens in the editor. The top bar has Undo and Redo, Transcript, Voice, Run again and Export. On the video itself, Hook and Subtitles open their settings.
Hook
A hook is a line of text shown over the start of the video to stop the scroll. Type the text, then pick its style, position, size, how long it stays on screen and how it comes in. It works the same as the hook in Viral Video Clips.
Subtitles
Change the caption style, font, colors and word highlighting, with a live preview on the video. See the Subtitle Generator docs for every setting.
Transcript
Open Transcript to see every word with its timing. Fix a word the captions got wrong, highlight a word, or give it an animation.
Timeline and Effects
The timeline under the video shows a filmstrip and the audio waveform. Zoom in with the mouse wheel or a pinch, and edges snap into place as you drag.
- Trim - Drag the clip start or end to cut the beginning or end of the video.
- Add an effect - Click the Zoom or Color lane where you want it, then pick Punch in, Color pop, Black and white, Flash or Vignette.
- Change an effect - Drag it to move it or drag its edges to resize it. Click it to type exact start and end times or to Remove effect.
Voice
Open Voice to see the voice, model and language used for the video. To make the same avatar say something else or speak in another voice, pick a voice, edit the lines, and click New video. You get a new project with the same photo, and the original stays as it is.
Run Again
Run again makes the video again from the same photo and speech, for example at a different resolution or for a new take. The new video replaces the old one, and your effects and subtitle style are kept.
Keyboard Shortcuts
| Key | What it does |
|---|---|
| Space | Play or pause |
| Left or Right arrow | Step one frame |
| Shift with Left or Right arrow | Step one second |
| Home or End | Jump to the start or end |
| Ctrl+Z | Undo |
| Ctrl+Y or Ctrl+Shift+Z | Redo |
On a Mac, use Cmd in place of Ctrl.
Export
Click Export to save the finished video with its captions, hook and effects.
- Desktop app - Choose the Quality (Original, High, Medium or Small), the Resolution and the Format (MP4, MOV or MKV).
- VCP Cloud - The video downloads at its original size, or at 4K on paid plans. When the export has captions, a hook, effects or a trim, it shows the estimated cost before it starts.
The export keeps going if you leave the screen. Come back to the project to see its progress or get the file.
4K Export
4K export upscales the video to 4K with FlashVSR. On the desktop app, 4K shows up under Resolution on NVIDIA RTX 30 series or newer graphics cards with 8 GB or more. On VCP Cloud, 4K is on paid plans.
VCP Cloud free plan
Free plans make 3 avatar videos a month, up to 10 seconds each, with a watermark. Paid plans add 1080p and 4K.
Use Cases
Social Media Videos
Make a talking video for TikTok, Reels or Shorts from one photo, with captions already on it.
Spokesperson Clips
Give a product, course or channel a face that can say new lines whenever you need them.
Ads and Product Demos
Make several takes of the same pitch in different voices or languages, then keep the one that works.
Characters and Stories
Bring a character portrait to life with a voice you designed for it.
Translated Versions
Make the same avatar say a line in another language with New video, so one face covers several markets.
Next Steps
- Create an Avatar - Make your own avatar and the places they appear in
- Text to Speech - Write and voice the lines your avatar will say
- Voice Cloning - Clone a voice to use in your videos
- Voice Design - Describe a new voice for a character
- Subtitle Generator - Every caption style and setting
Overdub
Select words in the AI Cut transcript, type new ones, and they are spoken in the speaker's own voice, with lip sync to match the mouth.
Create an Avatar
Make your own avatar for talking videos. Pick a place, a framing and a style, describe the person and their clothes, and choose from the portraits you get back. Make places from a description or a photo of a real room.