Restaurant daily special: avatar clip with the dish as product image

For a daily special, send the dish photo as product_image so the avatar presents it. At Sume's listed rates with a product image, 15 seconds costs about $2.91 on Standard, $3.87 on Plus and $8.70 on Max. The product image adds a small per-second amount over the no-product rate.

The script and the call

Write about 15 seconds: the dish, one fact about it, the price, and until when. Keep it to one dish per clip.

Call POST /v1/avatar-1.0/talking-video with an avatar_handle, the script, product_image set to a public HTTPS URL of the dish, and aspect_ratio 9:16. The URL must be fetchable: Sume rejects localhost, private-network and non-image URLs before it submits (Media inputs).

What it costs

At 15 seconds with a product image: Sume reserves on its planned billable seconds, which can differ a little from the length of your script, so these are estimates.

Look before you render

Check on a preview that the dish reads clearly in the first frame. A dark or cluttered photo is the usual cause of a weak result, so shoot or pick the photo before you write the script.

Label it

Tell viewers the presenter is AI-generated. Sume's disclosure record post lists the job fields worth saving per clip, and authored caption cues can burn a disclosure line into the video.