In this guide, we'll show you how to create and use LiveAvatars, HeyGen's next-generation platform for real-time, two-way digital interactions. Whether you're building a virtual assistant, a digital trainer, or an interactive customer support agent, LiveAvatars bring a human touch to your content with natural lip-sync, expressions, and gestures. Demo LiveAvatars here, and then let's dive into making one of your own! Note: HeyGen has rebranded Interactive Avatars as LiveAvatars. The feature has moved from HeyGen Labs to its own dedicated platform at app.liveavatar.com. If you created Interactive Avatars previously, migration support is available directly from the HeyGen platform.
What's covered in this guide
- Creating a custom LiveAvatar
- How to submit your footage
Accessing LiveAvatars
LiveAvatar is now a separate platform from the main HeyGen app. To get started, go to app.liveavatar.com, you can log in with the same credentials you use for HeyGen. Once logged in, you can toggle between My Avatars, Public Avatars, and Examples.
You can either choose one of the provided public avatars and add a Knowledge Base (Context) to it, or create a fully custom avatar from scratch.
Plan & pricing overview
LiveAvatar plans are stand-alone subscriptions, separate from your regular HeyGen plan, and are designed for users focused on real-time avatar interactions.
| |
|---|
| Access to all public LiveAvatars; 2-minute session limit, 1 concurrent session |
| Custom LiveAvatars; 1,000 credits for $100 |
LiveAvatar uses its own credit system:
- Full Mode (HeyGen handles LLM, voice & avatar): 1 credit = 30 seconds of streaming (~$0.20/min)
- Lite Mode (you bring your own LLM & voice; HeyGen provides the avatar layer): 1 credit = 1 minute of streaming (~$0.10/min)
Custom LiveAvatars are available exclusively to paid users.
Creating a custom LiveAvatar
Before recording, read through the guidelines carefully. Note that previous footage recorded for Video Avatars cannot be reused, the requirements for LiveAvatars are unique.
Record a continuous, uncut video of at least 2 minutes, divided into 3 distinct parts in this order:
- Listening state – 15 seconds (silent) Look at the camera as if listening attentively. Use facial expressions — raised eyebrows, smiles — to convey active engagement.
- Talking state – 90 seconds Speak clearly and conversationally, as though talking directly to another person.
- Idling state – 15 seconds (silent) Maintain a neutral, calm expression. You may nod gently — no extra facial expressions.
Tips for success
Background & environment
- Use a clean, static background or a green screen.
- Avoid moving objects, changing light sources, or reflections.
- If using a green screen, keep it off the floor and position yourself at least 5 feet from the backdrop to avoid shadows.
Camera & framing
- Use a 1080p HD camera or better (4K is not currently supported due to latency limitations).
- Frame yourself chest-up. Full-body recordings are not supported.
- If your hands are in frame, keep them still and in a resting position.
Lighting
- Use soft, even lighting, avoid harsh shadows on your face.
- Place two lights on either side with diffusion for a balanced look.
- Optionally, add a backlight above and behind for subject-background separation.
Performance tips
- Maintain steady eye contact with the camera throughout.
- Avoid loud breathing, swaying, or unnecessary movement during pauses.
- Keep jewelry, glasses, and shiny clothing to a minimum.
- Consider doing two takes: one with minimal movement, and one with 4–5 slow, subtle gestures (kept below shoulder height).
- Avoid hand gestures above your chest or pointing gestures.
- Record in a spot without background noise or music.
How to submit your footage
Once you're happy with your recording:
Go to app.liveavatar.com and click New Avatar. Upload your video footage, or record directly via your webcam.
Submit for processing.
Processing times by plan:
When your LiveAvatar is ready, you'll find it in your Avatar Library on the LiveAvatar platform.
Non-human or 3D avatars
You can also create LiveAvatars from 3D characters, photo avatars, or non-human faces. The process requires submitting a 2-minute video produced in HeyGen's Editing Studio plus a consent video of a real person. Note that submissions will initially be auto-rejected, simply reach out to [email protected] with your rejected Avatar ID for manual approval.
Using your LiveAvatar
Set up a Knowledge Base (Context)
Your LiveAvatar is powered by a Large Language Model (LLM). To guide how it responds, you'll set up a Context (formerly called a Knowledge Base):
On the LiveAvatar platform, select your avatar.
Click the Context button.
Add instructions, tone preferences, and information your avatar should know.
You can also customize your avatar's voice and rename it.
Integration modes
Once configured, you can deploy your LiveAvatar in two ways:
- Full Mode — HeyGen manages everything (LLM, voice, avatar). Fastest to set up; plug in and launch.
- Lite Mode — You bring your own LLM and voice; HeyGen provides the avatar layer. Offers maximum flexibility and customization.
Deployment options
Your LiveAvatar can be:
- Shared via a dedicated webpage link
- Embedded directly into a website or app
- Connected to your own LLM (e.g., ChatGPT or a custom-trained model) via the LiveAvatar API or SDK
Recap
You've now got everything you need to create and deploy a LiveAvatar! With real-time responsiveness, natural expressions, multilingual support, and flexible integration options, your avatar can handle a wide range of tasks, from customer support to education to sales.
In this guide, you learned:
- How to access LiveAvatars on the new dedicated platform
- Plan and pricing options, including the credit system
- How to record and submit footage for a custom LiveAvatar
- Tips for high-quality results
- How to configure and deploy your LiveAvatar
We're looking forward to seeing what you create! If you share your work on social media, tag @HeyGen so we can celebrate it with you.
