HeyGen vs Synthesia: API, pricing, avatars and limits

HeyGen and Synthesia both turn a script into a video of an AI avatar speaking it, and both sell a separate real-time avatar product. They differ most in how the API is sold (HeyGen sells pay-as-you-go API credits in US dollars apart from its plans; Synthesia includes API access from its Creator plan and draws on plan credits), in how long one video can be, and in how a custom avatar of you is made.

Everything below comes from each vendor's own pages, read on 2026-09-28, and is listed as each vendor states it. Neither vendor is ranked here: check the pages linked in each row before you buy.

How do HeyGen and Synthesia compare?

How is API pricing different?

HeyGen publishes a dollar price per minute for each avatar engine and avatar type, from $0.60 to $7.20 per minute in its API pricing article, and you buy API credits separately from any plan. Synthesia sells monthly and yearly subscriptions, with no paying per video: the Video API needs a Creator plan or above, and API minutes come out of the plan's credit pool.

The details are in HeyGen API pricing and Synthesia API pricing.

How is a custom avatar made on each?

• HeyGen trains a digital twin from footage: one person, facing the camera, with clear speech throughout. The subject then records a consent statement through a link that is valid for 24 hours. HeyGen's API pricing article says creating a custom digital twin through the API is only for Enterprise API users.
• Synthesia recommends building a personal avatar from one photo, with an optional voice sample; if you skip it, the avatar uses a Synthesia voice. The older video route takes 1-5 minutes of footage, and the consent video must be recorded live by the same person.
• Both also make avatars from a single photo; Synthesia notes that a photo avatar's motion comes from a speech-driven model, not from recorded footage. What is an AI digital twin compares the avatar types side by side.

Which one should I pick?

• Pick HeyGen if you want API access without a subscription, a dollar price per minute, or 4K export.
• Pick Synthesia if you want API use to draw on the same plan credits your team uses in the app, or one video needs many scenes (up to 150, and 4 hours in total).
• Pick by the live product if you need an avatar that talks back: both sell one, priced separately. See real-time vs video avatar APIs.

Where does a per-second avatar API fit?

If you need short clips from your own backend rather than long videos from an editor, Sume's Avatar 1.0 is a third option: one talking video covers an estimated 4-60 seconds, at 720p, billed at $0.184/s standard, $0.245/s plus, $0.55/s max (no product image). Its facts are on Generate avatar video; Sume vs HeyGen and Sume vs Synthesia compare each one directly.