Table of Contents
What HeyGen Actually Does
HeyGen generates presenter-led AI avatar videos from a text script, and separately can translate and re-voice existing real video footage into other languages with lip-sync. Avatar 2.0 is the current generation model — in blind tests, only 3 of 20 participants correctly identified a HeyGen 2.0 avatar as AI on first watch.
Video Translation with lip-synced real footage
Unlike avatar generation, this takes an existing video of a real person speaking and produces a translated, lip-synced version in 40+ languages — the best implementation of this specific feature in the category.
Getting Started: Account Setup
Go to heygen.com and sign up for the free plan — 3 watermarked videos per month, enough to judge Avatar 2.0's realism on your own script before paying.
Creating Your First Avatar Video
How to Make an Avatar Video
- Click Create Video and select an avatar from the library, or a template
- Paste your script into the text panel
- Choose a voice (or let it match the avatar's default) and preview before rendering
- Click Generate — rendering typically completes within a few minutes
Video Translation
How to Translate an Existing Video
- Go to Video Translation and upload your source video (a real recording, not just a HeyGen avatar video)
- Select the target language from the 40+ supported
- HeyGen transcribes, translates, and generates lip-synced audio matching the translation
- Review the output — commercial-quality results are typical, though a manual script check is worthwhile for accuracy
Use Video Translation on your best existing footage, not new HeyGen-generated content
The feature's real value is localizing content you already have — an existing product demo or testimonial — into new markets without re-shooting.
Custom Avatars
How to Create a Custom Avatar
- Go to Avatars → Create Custom Avatar
- Follow the guided recording process, typically a short script read on camera in good lighting
- Submit for processing (takes longer than generating a standard video)
- Once approved, your custom avatar is available in the library for any future script
Using Templates
How to Use a Video Template
- Browse Templates by use case — product explainer, social ad, training module
- Select one and swap in your own script, branding, and avatar
- Templates handle pacing and structure automatically, faster than building a video from a blank canvas
Full Workflow: Localizing a Product Video
Here's a workflow for taking one video to multiple markets:
Produce or select your source video (avatar-generated or real footage)
Either works as input for the next translation step.
Run Video Translation for each target market
Generate localized versions in parallel rather than sequentially re-scripting for each language.
Review each translated version for script accuracy
Spot-check translation quality, especially for less common languages.
Publish per-market versions
One source video becomes multiple localized assets without additional filming.
HeyGen Pricing 2026
| Plan | Price | Includes |
|---|---|---|
| Free | $0 | 3 videos/month, watermarked |
| Creator | $29/mo | 15 min/month, no watermark |
| Business | $89/mo | 60 min/month, Video Translation |
Start free — 3 real videos is enough to judge Avatar 2.0's realism before upgrading.
Common Beginner Mistakes
- Never trying Video Translation. It's HeyGen's strongest, most differentiated feature and easy to overlook if you only explore avatar generation.
- Skipping the translated-script review. A quick check catches occasional translation issues before a video goes public.
- Recording a custom avatar in poor lighting. Setup quality directly affects how convincing every future video with that avatar looks.
- Writing scripts that read like text instead of speech. Long, complex sentences that work fine on a page often sound stilted when an avatar delivers them. Write shorter sentences and read the script aloud yourself before generating — anything awkward to say will sound worse coming from the avatar.
Matching Voice to Avatar
HeyGen lets you pair any supported voice with any avatar, which is useful but easy to get wrong if the pairing feels mismatched to viewers.
- Preview several voice options against your actual script rather than a generic sample line — tone that works on a demo sentence doesn't always fit your specific content
- Match voice pacing to your avatar's intended energy level — a fast, energetic voice paired with a formal corporate avatar can feel inconsistent
- For recurring content (a training series, a regular show format), lock in one voice-avatar pairing and reuse it — switching pairings between episodes breaks viewer familiarity
Getting the pairing right on the first video in a series is worth the extra preview time, since changing it midway through a multi-part project creates a visible inconsistency across the set.