Make an avatar video look professional: scene photo of your office

To make a Sume avatar video look professional, give it a real setting: send a scene of type photo with a public HTTPS image_url of your office or shop, or a scene prompt that describes one, then check the first frame in a preview before the full render. Framing and background do more for a professional look than another round of script edits.

The motivation is in HeyGen's The State of AI Avatars 2026, a September 2026 survey of 1,000+ small business owners: 53.9% of those who shelved a recorded video cited not looking professional enough. It is a vendor survey, but the lesson transfers. The fields below come from Sume's avatar video guide and Media inputs.

Which scene option should you pick?

Sume has three ways to set the scene, each with a different field. Pick by how much control you need.

How do you send a photo scene?

The photo must be a fetchable public HTTPS URL. Localhost, private-network URLs, non-HTTPS URLs, signed or private URLs and mismatched content types are rejected before generation is submitted, so a link from a private drive will not work. Use a plain, well-lit photo without people in it; the avatar is the person in the frame.

Create a preview first so you can see the composition. Then render from the preview id.

curl -X POST https://api.sume.com/v1/avatar-video-previews \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: office-scene-preview-001" \
  -d '{
    "avatar_handle": "shop_owner",
    "script": "Welcome to the shop. Here is what is new this week.",
    "aspect_ratio": "16:9",
    "scene": {"type": "photo", "image_url": "https://example.com/office.jpg"}
  }'

What else affects a professional look?

Aspect ratio is the next lever. Sume supports 1:1, 3:4, 9:16, 4:3 and 16:9, defaulting to 9:16, so choose 16:9 for a website or email and keep 9:16 for short-form feeds. Resolution is currently 720p, which is worth knowing before you plan a large display.

Quality is the other choice: standard is the fastest path, plus the default, max the highest tier with slower turnaround. Preview stills are tier-independent, so you can approve composition once and still change the final tier.

What can go wrong?

A multi-scene video expects its backgrounds to resolve to one shared scene, so do not plan a dramatic location change inside a single request. Later scene stills in a shared-scene preview are pose-anchored continuations of the first frame, which keeps the look steady but not varied.

Sume does not clean up a cluttered photo. If the office photo is busy, the result will be busy. Retake the photo, or use a prompt for a simpler setting.

Should you add a product image too?

Only if the product is the point. product_image is an optional reference, and omitting it gives a productless avatar video. The cost difference is documented in the catalog rates, so check a dry run rather than guessing. A clean scene photo and a clear script usually do more for a professional look than extra references.

Media fields must be fetchable public HTTPS URLs, and Sume does not need a separate asset upload step for normal Avatar Video requests.