Avatar Shots is HeyGen's most expressive video generation feature yet, built in collaboration with Seedance2. Simply describe a scene in natural language, choose your avatar(s), and the engine brings it to life- complete with camera movement, dynamic backgrounds, and AI-generated audio.

This is the first HeyGen feature where you can place up to 3 avatars in the same video simultaneously. Whether you're recreating a debate scene, or just something totally wild- Avatar Shots makes it possible.

How to Access Avatar Shots

There are two ways to get there:

From the HeyGen home screen, click the Avatar tab in the top left corner, then select Avatar Shots from the menu (look for the 🆕 badge).
From the new HeyGen lobby, navigate to Avatar Shots / Seedance2.

All of your completed Avatar Shots videos will be saved in the Avatar Shots section- not in the main Projects page- for easy access.

Creating a Video

Step 1- Choose Your Avatar(s)

Your avatar thumbnail will appear in the prompt area. You can add up to 3 avatars to a single video- Currently only video-based real-human looks are available, and usually just one look out of an avatar slot. We should allow more look selection soon.

To add more avatars, click the + button in the prompt area.
To remove an avatar, click the small grey X button on the top right corner of the avatar thumbnail.

Step 2- Write Your Prompt

Describe the scene you want to create. Think of it like giving a director's brief- the more vivid and specific, the better the result.

Good prompts include:

The setting and mood (e.g. 'neon-lit Tokyo street at night, rain, cinematic')
What the avatar(s) are doing (e.g. 'two presenters facing each other in a debate')
Camera movement (e.g. 'slow push-in to close-up', 'wide establishing shot panning right')
Audio description- since you cannot upload audio separately, describe the sound you want (e.g. 'epic orchestral score', 'ambient city sounds', 'upbeat electronic beat')

💡 Fun example: Make the guy here become Super Saiyan in a very cool and dramatic way — lightning crackling around him, golden aura, epic orchestral music building to a peak.

Step 3- Add Reference Media (Optional)

You can upload images or videos to guide the visual style or include specific elements in the generated video. References can come from:

Your previously generated HeyGen assets
Files uploaded directly from your device

Use references when you want a specific background style, a prop, a color palette, or a visual mood carried through the video.

Step 4- Choose Duration and Orientation

Use the controls at the bottom of the prompt panel to set:

Duration: 5, 10, or 15 seconds (15 seconds is the maximum for now)
Orientation: Portrait or Landscape
Enhance: This would improve your prompt automatically if we find we can optimize it a bit.

Step 5- Generate

When you're happy with your prompt and settings, click the up arrow (↑) to generate.

⚠️ Important- No edits after generation: Once you click generate, the video is final. Credits are used immediately and you cannot make changes to the output- even small ones like adjusting a sound or tweaking a detail. If the result isn't what you were looking for, you'll need to start a new generation with an updated prompt. If you need precise, scene-by-scene control, consider using HeyGen Studio instead to avoid spending credits on multiple attempts.

How much credits does Avatar Shots costs?

Avatar Shots / Seedance2 uses 4 premium credits per 1 seconds of video at 720p resolution, and 10 premium credits per 1 second of 1080p video. 4K video creation is not currently available on Seedance 2 inside HeyGen. A 15 second video using Seedance2 can cost either 60 or 100 premium credits, depending if you choose 720p or 1080p resolution.

We're hoping to reduce this cost soon so more users can enjoy this!
This feature is available for all paid HeyGen plans.

Prompting Tips for Avatar Shots

Avatar Shots is powered by Seedance2, which has its own strengths compared to other engines. Here's how to get the best results:

Structure your scenes like a storyboard: 'Start wide on both presenters → cut to close-up of left avatar → pull back to full scene.' Seedance2 handles multi-shot prompts extremely well. Think in shots, not sentences.

Phrases like 'slow dolly-in,' 'crane shot overhead,' or 'tracking shot from the right' give the engine clear direction and it follows them with high consistency. Describe camera movement explicitly.

Since you can't upload audio, describe what you want- 'ambient coffee shop sounds,' 'cinematic orchestral swell,' or 'upbeat lo-fi music.' The engine will generate it. Include your audio in the prompt.

Generic prompts produce generic videos. Specifics like 'golden hour light through venetian blinds,' 'dust particles in the air,' or 'lightning crackling in slow motion' make a real difference. Go bold and specific.

Adding a style reference helps: 'cinematic, film-grade,' 'music video style,' 'documentary feel,' 'anime aesthetic,' or 'commercial product shoot.' Name the vibe.

Because each generation uses credits with no undo, spend time refining your prompt before hitting generate. The better your brief, the fewer attempts you'll need. Plan before generating.

Current Limitations

Maximum video length is 15 seconds.
Maximum 3 avatars per video.
Audio cannot be uploaded- describe the audio you want in your prompt instead.
Videos cannot be edited after generation- a new prompt is required for any changes.
Videos are saved in Avatar Shots, not the main Projects page.
Stable languages: English, Chinese (incl. some dialects), Spanish, Japanese, Korean, Indonesian
Also works: French, Italian, Turkish are also workable, but best to verify with a native speaker.