Knitwear sells on touch, and online there is no touch. The photo has to do the work: the shopper needs to see the stitch, how thick the knit is, how the sweater falls over the hips and how the sleeves sit at the wrist. A flat lay shows the stitch but not the fit, and an on-model shoot for every color of every sweater is expensive. This guide covers how AI knitwear product photography works in NotShot, what photo to upload, what to check before you publish, and what it costs.
Three things make sweaters and cardigans tricky, whether the photo comes from a studio or from AI:
For what a traditional shoot costs per image, see what catalog photography actually costs a DTC brand.
You upload one photo of the sweater or cardigan on its own. A flat lay, a ghost mannequin image or an existing packshot all work, as long as the whole garment is visible. For knitwear a few habits matter more than usual:
The full step-by-step version is in flat lay to model photo with AI: a step-by-step guide. If you shoot ghost mannequin images today, see the ghost mannequin alternative guide.
Knitwear is rarely worn alone in a photo. Up to 8 garments go into one 4K render, so a sweater can be shown with the trousers, the coat over it and the shoes it is merchandised with, in a single pass. Upload each piece separately and combine them in the render. For layering under outerwear, see AI jacket and coat product photography on model, and for the trousers, AI denim model photography or AI pants product photography. For the shirt underneath, see AI shirt product photography.
For the model, you can upload a photo of a model your brand already works with, and the identity stays the same across every render, or you create a brand-owned AI model once and reuse it across the whole collection. See uploading your model to NotShot.
Every render goes through an automated check before you see it. The check compares the render with the garment you uploaded: does the garment on the model match your photo, down to the pattern, neckline and hem; did the pose come out as requested; has the model's identity drifted; is there an artifact that would be obvious at full zoom. If a render falls under the bar, your credits come back automatically. How the judge works is covered in how the NotShot AI judge and auto refund work.
The check is automated and it is not infallible. For knitwear, zoom in on the stitch, the cuffs and the hem band at full resolution before you publish. Every render is delivered at 4K, which keeps fabric texture sharp when a shopper zooms in (see why the 4K upscaling step matters). If the stitch looks soft on a render you otherwise like, the texture and detail pass (2 credits) sharpens fabric texture and fine detail without starting over.
A shipped 4K render is 4 credits. At the pay-as-you-go rate of about $0.50 per credit that is about $2.00 per image. On a subscription it drops: the Bundle plan is $29 a month for 100 credits, about $1.16 per render, and the Studio plan at $299 for 1,500 credits brings it to about $0.80. A knitwear drop of 10 sweaters in 3 colors each, with one on-model image per color, is 30 renders, 120 credits, roughly $60 at the pay-as-you-go rate. Renders that fail the check are refunded. New accounts start with 50 free credits and no card is needed. For the full breakdown, see NotShot credit cost explained.
There is also a short overview for sweatshirts and hoodies on the sweatshirts page.
