HeyGen Video Agent: what it does, its API, and pricing
HeyGen's Video Agent turns a text prompt into a finished avatar video, in the app or via POST /v3/video-agents. Its inputs, outputs, and API price.

HeyGen Video Agent is HeyGen's prompt-to-video agent: you describe the video in text, and the agent handles scripting, avatar selection, scene composition, and rendering. Developers call it with POST /v3/video-agents, and over the API HeyGen lists Video Agent at $5.82 per minute of video, charged by the actual seconds generated.
Every HeyGen fact here comes from HeyGen's own Prompt to Video docs, its API pricing article, and its pricing page, all read on 2026-09-28. Check them again before you buy; prices change.
What does HeyGen Video Agent take and return?
A prompt is the only required field. The agent chooses the avatar and voice unless you pass avatar_id or voice_id, and it detects orientation from the content unless you set it.
| Item | What HeyGen's docs say |
|---|---|
| Endpoint | POST /v3/video-agents |
prompt | Required, 1–10,000 characters. |
mode | generate (default), "one-shot, fire-and-forget"; chat, multi-turn, where the agent pauses for review and supports revisions. |
orientation | landscape or portrait; auto-detected if omitted. |
files | Up to 20: images (png, jpeg), video (mp4, webm), audio (mp3, wav), and pdf. |
| Other options | style_id, brand_kit_id, brand_glossary_id, callback_url. |
| Session statuses | thinking, generating, completed, failed. |
| Finished video | video_url, a presigned download URL present when completed. The response also has captioned_video_url and an SRT subtitle_url field. |
How do you call the HeyGen Video Agent API?
Create a session, then poll twice: GET /v3/video-agents/{session_id} until a video_id is assigned, then GET /v3/videos/{video_id} until the video is completed or failed. Or pass a callback_url and HeyGen POSTs when the video completes or fails. HeyGen's own curl example:
curl -X POST "https://api.heygen.com/v3/video-agents" \
-H "X-Api-Key: $HEYGEN_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"prompt": "Create a 45-second explainer about our Q3 product launch. Use a friendly, upbeat tone. Include the attached slides as visual context.",
"orientation": "landscape",
"files": [
{ "type": "url", "url": "https://example.com/q3-launch-deck.pdf" }
]
}'How much does HeyGen Video Agent cost?
It depends on where you use it. HeyGen prices the API and the app separately:
- API: pay-as-you-go credits bought in US dollars. Video Agent (Prompt to Video) is $5.82 per minute, or 116 credits per minute, charged by the actual seconds generated. HeyGen says the table applies to videos generated via API only.
- API credits: HeyGen API pricing covers how the credits are bought, when they expire, and the concurrency limit.
- App: the Free plan includes up to 3 videos per month and limited trial access to Video Agent. Paid plans (Creator at $29/month, Pro from $49/month) come with monthly credits, and the pricing page says the credits a video uses vary with the model, duration, and complexity. It does not list a per-minute Video Agent rate for the app.
What does a video agent API look like without an avatar?
HeyGen's docs describe the endpoint as creating AI avatar videos. If your brief doesn't need a presenter, Sume's Agent Completions take the same kind of brief: POST /v1/agent/completions with an instruction, caller data in input, up to 30 images, and a required generation_spend_cap_usd. It answers 202 with an async run receipt, not choices[], and you poll it. Saved recipes called Formats, including a first-party catalog at the sume handle, return the same receipt shape. Generation is metered at the API pricing rates, plus a 5.5% agent fee by default. HeyGen alternatives with an API covers avatar APIs side by side.
Sources
Related posts
More in Agents
- Image generation MCP server: how Sume's generate_image works
Sume's hosted MCP server has a paid generate_image tool: a prompt in, a job id back in milliseconds, then jobs_wait and jobs_result for the images.
- MCP vs function calling: how they differ and fit together
Function calling lets a model ask your app to run a function you defined; MCP puts tools on a server any client can discover. How they fit together.
- MCP vs REST API: what's the difference and when to use each
A REST API is endpoints your code calls; MCP lets an AI app discover and call a server's tools at runtime. How they differ, and when to use each.
- PDF to video AI: how to turn a document into a video
PDF to video AI turns a document's text and figures into a narrated video. On Sume today, you extract the text and export the figures as images first.
Written by Sume