heygen/avatar-4

Avatar 4 turns a photo into a talking avatar that speaks your text or lip-syncs to an audio file.

$0.10 / second
GPU: H100

Model Inputs

file

Photo to animate (.jpg/.png/.webp). It must contain a clear face

string

The text the avatar will speak

string

Name of the voice to use for the avatar

file

Optional speech for the avatar to lip-sync to (.mp3/.wav/.flac/.m4a/.ogg). When set, it overrides prompt and voice

string

Talking style - 'stable' for minimal movement, 'expressive' for more animation

string

Facial expression for the avatar to hold, e.g. friendly, serious, excited

string

Video resolution preset. Options: 360p, 480p, 540p, 720p, 1080p

string

Aspect ratio of the output video. Supported values: '16:9', '9:16', '4:5', '5:4', '1:1', and 'auto'. 'auto' preserves the source aspect ratio when HeyGen can read it, falling back to '16:9' otherwise.

boolean

Whether to add captions to the video

Result

Idle
Waiting for your input...

Pricing

heygen/

avatar-4

You are billed for what this model actually produces — by the second — so a shorter request costs less and you never pay for idle time.

Rate
$0.10 / second
or
10 seconds / $1

For example, a request of 60 seconds costs about $6.00.

You are charged for what the request actually produces, so the final amount scales with the seconds you use.

ProviderPrice ($)Saving (%)
Synexa$0.10-
fal$1.0090.0%

Comparison prices are for an equivalent request.

Readme

Avatar 4 animates a photo of a face into a speaking avatar. Drive it either with prompt (the text to speak, using a named voice) or with audio_url - supplying audio overrides both prompt and voice.

talking_style chooses between stable (minimal movement) and expressive (more animation); expression sets the facial expression and caption burns in subtitles. Resolution runs from 360p to 1080p, and aspect_ratio: auto preserves the source shape.

Custom backgrounds exist upstream but are not exposed here, because they take a nested object payload.