Real Estate Listing Reel From 8 Photos: A $5.23 Workflow

The short answer

You can turn eight listing photos into a 40-second vertical tour for about $5.23 on Sume list prices: eight 5-second image-to-video clips on Gemini Omni Flash 1.1 at 720p ($5.00), one generated music bed ($0.125) and one timeline render of 40 seconds, which bills as one minute ($0.10). Prices are the catalog values the Sume team recorded on 2026-10-08, and they are Sume's price, which is the provider list price times 1.25.

Why Omni Flash fits a photo tour

Gemini Omni Flash 1.1 (gemini-omni-flash-1.1) takes 3 to 10 seconds, 16:9 or 9:16, and accepts an image_url for image-to-video through the Video Router. A single room photo moving for 5 seconds is a normal fit for that envelope. Native audio is always on for this model and the API rejects generate_audio: false, so plan to replace the clip audio with the music spine in the timeline step.

Steps

Send each photo to POST /v1/video-router/generate with model set to gemini-omni-flash-1.1, image_url, duration 5, resolution 720p and aspect_ratio 9:16. Give every call its own Idempotency-Key, so a client retry cannot create a second paid clip. Put the camera move in the prompt (slow push in, no people added) and keep it the same across rooms so the cuts feel related.

Ask the Music Router for a bed with the length in the prompt, because the API rejects a duration field. Then call POST /v1/timeline-1.0/render with audio.url set to the music file, audio.duration_seconds 40, and eight video[] slots of 5 seconds each. Use a fade transition of 0.25 seconds on slots after the first. Output defaults to 1080x1920.

• Timeline sources must already be Sume-hosted: import files first with POST /v1/media-imports.
• POST /v1/timeline-1.0/plan is an unbilled preflight that returns billable_minutes and estimated_cost_usd_micros before you pay.
• Do not add rooms or features the photos do not show; a generated clip can invent detail, so review each clip against the listing.

What changes the price

Going to 1080p moves Omni Flash to $0.1875 per second, so the same 40 seconds of clips costs $7.50 instead of $5.00. A 10-second hero shot for the kitchen adds 5 seconds, or $0.625 at 720p. The timeline step stays at one minute until you pass 60 seconds of output.

If you only need still-image motion for a feed ad, 3-second clips are the minimum Omni accepts: 8 x 3 s at 720p is $3.00.

Limits to plan around

Omni Flash 1.1 gives you only 3 to 10 seconds per clip, so a tour is always a sequence of short clips, not one continuous walkthrough. It will not know the floor plan. If a buyer could read the reel as a statement of the layout, order the rooms the way you would walk the property and say in the listing text that the video is illustrative.

Timeline 1.0 accepts up to 1,800 seconds of spine and up to 200 slots, so a longer tour with 24 rooms and a 2-minute spine is possible; the render line then becomes $0.20. The rest of the cost grows with the number of clips, because video is billed per output second.