Instacart Immersive Feed: 9:16, 5 to 30 s, a recipe clip from a photo
Instacart's Immersive Feed takes 9:16 video of 5 to 30 seconds and is in early pilot. Prepare a recipe-style clip from a product photo with the Sume video API.

Instacart's Immersive Feed takes vertical 9:16 video between 5 and 30 seconds, and Instacart describes it as an early pilot with select brand partners. If you are a grocery or food brand and expect to get in, prepare a short recipe-style clip now. You can make one from a single pack shot with the Sume video API, then keep it ready as an existing vertical asset.
What does Instacart say?
Because it is a pilot, check with Instacart whether your brand can book it.
| Topic | What the release says |
|---|---|
| Format | "9:16 vertical video between 5 and 30 seconds in length" |
| Status | "currently in early pilot with select brand partners"; more partners planned later this year |
| Booking | Through Instacart's Ads Manager, using existing vertical video assets or creating new ones |
| Shopper path | Browse the feed in retailer storefronts on Instacart and add items to the cart |
How do I make a clip that fits?
Use POST /v1/videos with your pack shot as a first_frame, aspect_ratio: "9:16" and a duration inside both limits. seedance-2.5 accepts 4 to 30 seconds, so ask for 5 or more to stay inside Instacart's range. Confirm what your chosen model lists with GET /v1/videos/models.
Write the prompt as one simple cooking moment, such as stirring a jar of sauce into a pan. Ask for no text on screen, because the feed is where the shopper sees the product and the add-to-cart button, not your caption.
{
"model": "seedance-2.5",
"prompt": "Overhead kitchen shot: tomato sauce from this jar is poured into a pan of pasta and stirred, steam rising. No text.",
"aspect_ratio": "9:16",
"resolution": "720p",
"duration": 12,
"frame_images": [{
"type": "image_url",
"image_url": { "url": "https://cdn.example.com/sauce-jar.jpg" },
"frame_type": "first_frame"
}]
}What do I check before booking?
Check the length with video inspect (frames: false returns the probe only). Check that the label on the jar is the same as on the shelf. A food clip that shows the wrong pack, or a food state that the product cannot produce, will not help a shopper.
Which clips are worth preparing?
Prepare more than one, since the feed is a place where a shopper browses and then adds items to a cart.
- A 6-second pour or stir shot of the product.
- A 15-second recipe in two steps.
- A 25-second version with a second serving idea.
- A still of the finished dish, made with the Image API, for any slot that takes an image.
Sources
Related posts
More in Use cases
- Job opening video with an AI host: 16:9 and 9:16 versions
A 30-second hiring video with a Sume avatar costs $5.52 on Standard or $7.35 on Plus per aspect ratio. A 16:9 and a 9:16 cut on Plus is $14.70.
- Kids story, five 8-second scenes: Omni $5.00 vs H3 $3.00
Five 8-second scenes cost $5.00 on Gemini Omni Flash 1.1 at 720p, $3.00 on MiniMax H3 at 768p and $4.00 on H3 Max, priced on Sume. Plus a reference body.
- Localize One Ad to Korean: Burned-In Captions With Hangul Styles
Add Korean captions to an ad: why Latin styles return a 400, which Hangul style to pick, how a restyle skips a second transcription, and what a job costs.
- Korean restaurant reel captions in Gmarket Sans: one font field, $0.20
Burn Korean captions on a restaurant reel in Gmarket Sans: set font gmarket-sans, style black-outline, language ko. $0.20 for up to 60 seconds.
Written by Sume