Airbnb house-rules video: an avatar in front of your living room

A short-term rental host can make a house-rules video by sending a script and a photo scene to POST /v1/avatar-1.0/talking-video. The scene field takes { "type": "photo", "image_url": "https://..." }, so the avatar appears in front of your own living room photo, and the output is 9:16 by default.

The result is a friendly welcome that you can send in a booking message, not a replacement for the written rules the platform shows guests.

Why a short video helps

Guests skim written rules. A 30-second clip covers check-in, quiet hours and trash day in the voice of one presenter. Trend roundups list avatars as common in support and localized sales material (AI Video Generation Trends, read 2026-10-07), and a host's welcome is a small instance of support content.

Sume accepts scripts it estimates at 4 to 60 seconds. Keep to the few rules guests most often break.

The request

Create the avatar once with POST /v1/avatar-1.0/generate, using a prompt, a profile or a reference image, and reuse its handle. For each unit, change the scene.image_url to that unit's photo. Media fields must be fetchable public HTTPS URLs.

Use a clear photo of the actual room, not a stock image, so the video does not promise a space the guest will not find.

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: unit-12-house-rules-001" \
  -d '{
    "avatar_handle": "host_avatar",
    "aspect_ratio": "9:16",
    "quality": "standard",
    "scene": { "type": "photo", "image_url": "https://example.com/unit-12-living-room.jpg" },
    "script": "Welcome. Quiet hours start at ten. Please take your shoes off at the door, and the trash goes out on Tuesday. Message me any time.",
    "captions": { "enabled": true, "style": "slam", "language": "auto" }
  }'

Settings and limits

Resolution is 720p at this time. Quality accepts standard, plus (default) and max, and standard is the fastest path. Inline captions are burned in after generation and rejected for scripts estimated at over 60 seconds. If you want to review first frames before paying for a full render, use an avatar video preview and call generate-video on the preview id.

Be clear it is an avatar

Tell guests the host presenter is an AI avatar, and keep the real contact details in the listing. If a platform rule or local law says rules must be given in writing, the video is an addition, not the notice.