Does a product image cost extra in a Sume avatar video?

Yes. Passing a product_image to Sume's avatar talking-video route uses a higher per-second rate: $0.194 instead of $0.184 on standard, $0.258 instead of $0.245 on plus, and $0.58 instead of $0.55 on max. The difference is $0.010, $0.013 and $0.030 per second. Omit product_image and you get a productless avatar video at the lower rate.

These come from the catalog rate card sume_avatar_video_fixed_tier_rates_2026_07_02. Check GET /v1/catalog for the live number before you set a budget.

The rates and a 30-second clip

Thirty seconds, with and without the product image.

When the extra is worth it

The product image gives Sume the product reference for the clip. For a spoken product pitch it is the point of the clip, and the premium is under a dollar on a 30-second max render. For a talking-head update, an intro, or a how-to where the product never appears, it is cost without benefit.

product_image must be a fetchable public HTTPS URL. Localhost, private network addresses, non-HTTPS URLs and non-image responses are rejected before generation, so host the file where Sume can reach it.

curl -X POST https://api.sume.com/v1/avatar-1.0/talking-video \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: mug-pitch-001" \
  -d '{
    "avatar_handle": "studio_presenter",
    "script": "This mug keeps coffee hot for six hours.",
    "product_image": "https://example.com/mug.png",
    "scene": {"type": "prompt", "prompt": "Bright kitchen counter"},
    "quality": "plus"
  }'

Cheaper ways to test

• Create an avatar video preview with the product image and review the first frame before the full render.
• Run the first draft on standard; the preview stills are reused if you later render at a higher tier.
• If the product only shows in one scene, consider a separate clip for that scene rather than paying the higher rate on the whole video.

Limits

The product rate applies to the whole video, not just the seconds when the product is visible, so planning the clip matters. Use a clean, well-lit product photo, and review the first frame with a preview before you scale up. Rates can change; the numbers above are from the rate card named in the first section. To decide for a whole program, multiply the extra by the seconds you expect to render in a month: 2,000 seconds of plus with a product is $26 more than without, which is small next to the production cost of the product photo itself, and large only on max at volume. Put the choice in your request builder as a flag so editors do not attach a product image by habit.