HeyGen vs Argil: avatar APIs, avatar creation and pricing

HeyGen and Argil both turn a script into a talking-avatar video through an API; they differ in who can use the API and how it bills. HeyGen sells API access as pay-as-you-go credits in US dollars that any user can buy without a plan, priced per minute by avatar engine and type. Argil's API prices apply to Classic plans and above and draw on the same credits as its app: 160 credits per minute of video, 20 per minute of voice, plus a royalty on Argil's own v1 avatars.

Every vendor fact below comes from HeyGen's and Argil's own pages, read on 2026-09-28 and listed under Sources. Neither is ranked here.

How do the HeyGen and Argil APIs compare?

How does pricing differ between HeyGen and Argil?

HeyGen prices the API in dollars. The rate depends on the engine and avatar type: on the Avatar IV engine, a photo avatar costs $2.31 a minute and a digital twin $4.83, charged by the seconds generated. Pay-as-you-go credits expire after 12 months.

Argil prices the API in credits shared with its app, so the dollar cost depends on what your plan pays for credits. The avatar royalty applies only to Argil's avatars: if you train your own avatar, you do not pay it. Accounts over 60,000 credits a month are routed to Argil's enterprise contact.

How do you create a custom avatar on each?

Both create avatars from a photo, and both have a footage-based path, but Argil is retiring its footage path.

One caveat on HeyGen: its developer docs show a digital twin created with POST /v3/avatars, but its API pricing article says that is only available to Enterprise API users (HeyGen API pricing). On Argil, the quickstart assumes you have trained at least one avatar in the app before your first API video.

• HeyGen photo avatar: POST /v3/avatars with "type": "photo" and one still image. There is no consent step in the API, but HeyGen says to have the person's permission first.
• HeyGen digital twin: "type": "digital_twin" trains on footage of the person (HeyGen's target is 15 to 600 seconds) and clones the voice from it. A consent link, valid for 24 hours, lets the subject record a statement on camera; uploading a consent video instead is Enterprise-only.
• Argil image avatar: POST /avatars with "type": "IMAGE": PNG, JPEG or WEBP, exactly 16:9 or 9:16, 720p to 4K, 10MB at most. It costs 2 credits and typically becomes ready within 30 seconds.
• Argil video avatar: deprecated. It needs a 1-5 minute dataset video and a consent video of 30 seconds or less, and Argil says it will be removed in a future version.

Which one should I pick?

• Pick HeyGen if you want API billing in dollars without a subscription, or a choice of avatar engines at different per-minute rates.
• Pick Argil if you already use its app on a Classic plan or above and want API videos to draw on the same credits, with avatars built from one image.
• Sume's Avatar 1.0 is a third option for script-to-video: one job takes a script Sume estimates at 4-60 seconds (Generate avatar video). Sume vs HeyGen and Sume vs Argil compare it with each.