Can HeyGen use my avatar and voice? What its terms say
On Creator, Pro and Business plans you grant HeyGen an irrevocable license to your content, including to train its models. Ask to opt out by email.
Yes, HeyGen can use your avatar and voice recordings under its terms. On Creator, Pro and Business plans you keep ownership, but you grant HeyGen a royalty-free, transferable, sublicensable, worldwide and irrevocable license to your content, including to train its AI models. Its privacy policy says it may use input to train its models, including to make its avatar creation models more accurate, and you can ask to opt out of that training by email.
Everything about HeyGen here is quoted from its own Terms of Service (last updated July 23, 2026) and Privacy Policy (last updated August 11, 2026), both read on 2026-09-29. This reports what the pages say. It is not legal advice, and it doesn't say what "irrevocable" means for you in a dispute.
What does HeyGen's license to my content say?
The license sits in section 3 of the Terms, which covers users on the Creator, Pro and Business plans. "Your Content" there means your input (the footage, voice, scripts and images you upload) and the output HeyGen generates for you. The table quotes the parts that decide what HeyGen may do with an avatar or a cloned voice:
| Question | What HeyGen's page says | Where |
|---|---|---|
| Who owns what you upload? | HeyGen "does not claim to own any of Your Content" | Terms, section 3 |
| What license do you grant? | To "access, use, host, cache, store, reproduce, transmit, display, publish, distribute, and modify" your content | Terms, section 3 |
| For what purposes? | "to operate, improve, promote and provide the Services and to develop new services and products, including to train" its AI and machine learning models | Terms, section 3 |
| On what terms? | "royalty-free, transferable, sublicensable, worldwide and irrevocable", and the section survives termination | Terms, section 3 |
| Is voice and face input covered? | Input includes "text, voice, scripts, images, videos"; face imagery is processed to "create and personalize avatars" | Privacy Policy, section 3 |
| Is it used for training? | To train its models, "including to make our avatar creation models more accurate" | Privacy Policy, section 4 |
| Can you opt out of training? | You "may request to opt-out" by contacting HeyGen by email | Privacy Policy, section 4 |
| Biometric data | In its GDPR section: explicit consent is obtained during avatar creation; withdrawing it may prevent avatar creation | Privacy Policy, GDPR section |
Can HeyGen use my voice to train its AI?
The privacy policy lists voice among the input it collects, and says it may use input "to train and enhance the models that power our Services". Among the rights you "may have" depending on where you live, it lists a right to object, including to your information "being used to train our models".
- Ask for a training opt-out by email; the address is on the privacy policy.
- Where HeyGen relies on biometric consent under the GDPR, you can withdraw it by email. HeyGen says withdrawal may prevent you from using avatar creation.
- To delete, the privacy policy says you can delete data from your account or delete the account, and that HeyGen strives to delete within 72 hours of a request, unless the law requires or permits it to keep the data.
Is the Free plan different?
Yes, in part. Section 4 of the Terms covers the Free plan. It says you own your input, but you get only a limited license to the output "solely for personal, non-commercial, and internal evaluation purposes". The input-license paragraph quoted above appears in section 3, the paid-plan section; the privacy policy's training purpose does not name a plan.
What should I check in any avatar tool's terms?
Before you record a digital twin anywhere, read the content-license paragraph and the privacy policy's purposes list. Four words matter most: who "owns" the input, whether the license is "irrevocable", whether it covers "training", and whether it survives when you close the account.
For comparison, Sume's Terms of Service (last updated August 10, 2026) say you retain ownership of "voice input" and "likenesses", and grant Sume the rights needed to process your content "to provide, secure, support, and improve the Services". The terms also require you to have permission to use "any person's likeness or voice". Neither Sume's Terms nor its Privacy Policy states whether customer content is used to train models. Sume's API creates an avatar from a text prompt, structured traits or a reference image (Create new avatar). For what a twin is, see What is an AI digital twin?; for making one of yourself, How to clone yourself with AI.
Sources
Related posts
More in Sume Avatar 1.0
- Creatify vs Synthesia: ad avatars or training videos, compared
Creatify sells avatar video ads on credit plans with separate API plans; Synthesia sells presenter videos in 160+ languages with API from Creator.
- Does AI keep your photos and face? What avatar tools store
If an AI tool makes a reusable avatar, it keeps your face on purpose. How long HeyGen, Synthesia and Sume keep it, and how to get it deleted.
- Does HeyGen do captions? What its video API returns
Yes. HeyGen's v3 video API returns an SRT subtitle file when you set caption, and burns captions into the video when you also set a style.
- Does Higgsfield do AI UGC? Its UGC tools and API, explained
Yes: Higgsfield's app has Lipsync Studio and Click to Ad for UGC-style clips. Its API docs list video models, not a lip-sync or UGC ad endpoint.
Written by Sume