Cleaning service before-after reel: still above, clip below, 2 cents

A cleaning company can show a before photo above a finished-room video in one frame with Sume timeline compose, and the compose job costs a flat $0.02. The result is one MP4 where the still stays on screen the whole time and the clip plays underneath, which is the split-screen before-and-after layout customers already recognize.

This post is about that one tool and its limits. Compose does not generate anything. It takes one still and one video, both already hosted on Sume, and tiles them.

How compose lays out the frame

The compose docs define two operations. stack tiles the still and the video into two regions; the defaults are a horizontal split with the still on top at a ratio of 0.5. overlay keeps the video full-frame and pins the still on top of it as a plate, with a width ratio between 0.05 and 1.

For a cleaning reel use stack, with the before photo on top and the clip below, or switch to a vertical split for a landscape frame. The output length comes from the video layer, up to 300 seconds, and the still can never make the clip longer.

• Field that picks the layout is operation, not mode.
• Mixing stack keys with overlay keys returns a 400.
• The default mode is async; pass mode: "sync" to wait up to 30 seconds.

Where the files must live

Both URLs must already be this workspace's media.sume.com artifact or asset, and the still must probe as a still. A phone photo on your own web host will not work. Files generated by Sume or stored in your workspace assets qualify. The one import endpoint, POST /v1/media-imports, only accepts public TikTok or Instagram video URLs at $0.15 each, so it will not pull in a camera-roll photo. Check the media inputs docs for how assets get into the workspace before you plan the shoot.

Cost of the whole reel

The finished-room clip is the main cost. A 6-second Wan 3.0 clip at 480p is $0.0625 x 6 = $0.375, charged as 38 cents per job, and compose adds $0.02. The reel is about $0.40 (read 2026-10-10, Video Generation and compose docs).

Ten jobs a week at that price is about $4.00.

Choosing stack or overlay

Stack is the right call when you want a true before and after, with the old state fixed on screen and the new state moving. Overlay suits a different story: the video fills the frame, such as footage of the finished building, and the still sits on it as a plate, for example a before photo or a price card, placed at the top, center or bottom with a width ratio you choose.

For a cleaning company, think in terms of the viewer's eye. A horizontal split with the before above the clip reads quickly on a phone held upright. A vertical split puts the two states side by side, which suits landscape placements such as a website banner. Either way the layout fields only accept their own vocabulary, so keep the request to the keys that match the operation you picked.

Because the job returns one MP4 shot, you can feed it into a longer Timeline render later with a voiceover and a closing card if you want a full 20-second spot. Compose on its own is enough for a social post.

• Stack, ratio 0.5: the default half-and-half.
• Stack, ratio 0.3: a thin before strip over a large result.
• Overlay, position bottom: full-frame result footage with the before photo as a small plate.

Using the before-after format

The public catalog under the handle sume includes a sume-before-after format. Read GET /v1/formats/sume/sume-before-after and run it once before you promise anything to a client, because the repo does not ship format descriptions and you should see its inputs and price for yourself. See the Format catalog.

One honest caution: generated finished-room clips show an imagined clean room, not your crew's work. Use them as a lead-in for the real before photo, and say in the caption which part is a photo.