AI Saree Photography for your catalogue

From the photo you have to the photo the listing needs

Your sarees arrive from the loom or the godown flat. Your listing needs them draped on a person, in light that shows the colour truthfully, with the border and pallu doing the selling. Between those two photos there used to be a studio booking, a model, a draping expert and a week of waiting. Now there is an upload.

A real pair: the image on the right was generated from the photo on the left. Same border, same lattice, same trim. This one input also produced every image in our pose guide.

A saree is two garments, and the tool knows it

A saree ships as six metres of drape fabric plus a blouse piece. Generic photo tools see one image and one garment, which is why their saree results disappoint: the blouse gets ignored or mangled, and the drape ignores how a pallu actually falls. SareeViz takes the saree and the blouse as separate uploads and drapes them as one outfit. In our analysis of 207,556 catalogue images, 69% of saree jobs used that separate blouse upload. It is not an edge case, it is how sarees work.

If you are weighing options first, our comparison of AI saree photography tools ranks seven tools including ours, and the saree photo editing guide walks through the whole process step by step. For drape styles specifically, see saree draping software. Selling intimates alongside your ethnic range? AI innerwear and lingerie photography covers that category, where booking a model is hardest. And if it is a whole season you are costing rather than one design, the catalogue photoshoot arithmetic works it out per design against a studio quote.

The drape decides what a buyer can judge

Six metres of fabric can be worn a hundred ways, and the way you choose decides which part of your saree a buyer ever sees. A drape is not styling. It is a claim about where your design lands on a body.

Then the practical part: think about the thumbnail. On a marketplace grid a buyer meets your saree at roughly the size of a stamp, cropped to the middle of the frame. Whatever the drape puts on the model’s back does not exist at that size. If your money is in the pallu, get the pallu to the front.

One honest note on what this tool does here. The drape it generates sits in the contemporary Nivi family, and the draping preview exists to tell the AI which of your uploaded designs belongs to the pallu, which to the pleats and which to the border, rather than to switch between regional styles. Style follows the pose you pick and the instruction you write. If showing one saree in three named regional drapes is the actual job, our comparison of saree draping software names the tool built for that, including one that is not ours.

Every fabric photographs differently

The tool copies what is in your photo, so the fabric decides how you shoot, not how you generate. Five behaviours cover almost every saree in an Indian catalogue.

• Silk, so Kanjeevaram, Banarasi, Paithani and tussar. Stiff enough to hold a sculpted pleat, and reflective enough to throw a hard highlight. In direct sun the fold nearest the light blows out to white and takes the colour with it. Shoot in open shade or by a window with the sun off the fabric, and the weave stays readable.
• Cotton and handloom. Matte, absorbs light, pleats crisply and photographs honestly. The risk is the opposite one: a flat, dull frame that undersells a good weave. This is the fabric where a close-up of the texture adds the most, because the weave is the product.
• Georgette and chiffon. No body at all when laid flat, so they read as a puddle of colour on a table. Everything worth paying for, the fall, the swing, the way pleats separate as the model moves, only exists on a person. This is the category where an on-model image beats a flat lay by the widest margin, and where a walking or turning pose beats a straight stand.
• Organza and tissue. Translucent and reflective at the same time. Whatever sits behind the fabric comes through it, so a dark table under an ivory tissue saree drags the colour grey, and the tool faithfully copies the grey. Shoot on plain white or light grey, never on wood or a patterned bedsheet.
• Zari, border and surface work. Metal thread only looks like metal when light reaches it at an angle. A straight-on phone flash flattens zari into a beige stripe, and a beige stripe is what gets reproduced. Photograph the border on its own, straight on, in daylight, and upload that as its own close-up so the work survives at listing size.

None of this is about the AI. It is about what your photograph contains, which is the one part of the job still in your hands.

What to photograph before you upload

Four photos cover almost every saree, and most sarees need only the first one.

1. The full saree, flat or hanging, cropped so only fabric is in the frame. This photo decides the colour, the fabric and the border in every output, so use your original product photo and never an image the tool generated earlier.
2. The blouse piece alone, cropped tight, with no saree fabric anywhere in the frame. A blouse photo with the border creeping into one edge is the commonest reason a blouse design turns up on the pleats.
3. The pallu alone, but only when the pallu carries a design the full-saree photo cannot show clearly. On an all-over print, skip it.
4. The border or the buti, straight on, only when the work is fine enough to be lost at full-saree scale.

The shooting rules are short. Daylight, no filter and no beautify mode, a plain white or light grey surface, the whole saree in frame, the file a couple of MB rather than fifteen. And stop at four: seven overlapping photos of one repeating print give the AI seven instructions to reconcile, and the result is worse than one good photo. The step-by-step, slot by slot, is in the saree photoshoot guide, and the cropping detail is in preparing your design photo.

What a marketplace wants from a saree listing

Myntra, Amazon and Flipkart each publish a minimum photo size that a 1K image cannot reach, and each asks for a different set of views. Meesho publishes no size rule at all. None of the four allow a logo or a watermark on a listing photo, which rules out the shop-name overlay that a lot of WhatsApp catalogue images carry.

Rather than ask you to hold those rules in your head, pick the marketplace in the app and the poses, background, model and output size are set to what that platform publishes, then locked so every photo in the set matches. The marketplace guide lists what each one asks for and exactly what the preset sets, with the full photo set generated from a single saree so you can see the difference before you spend anything.

What one saree costs to photograph

Images are priced in credits by resolution, and a credit is ₹10. That is ₹10 for a 1K image, ₹15 at 2K, ₹20 at 4K and ₹25 at 8K, before 18% GST. There is no subscription, so nothing runs in a month you do not shoot.

A marketplace preset makes a fixed set of photos per design, which turns that rate into a per-design price you can plan a season around:

So a forty-design season run as a Meesho ecommerce photoshoot is ₹1,200 before GST, and the same forty designs shot for Myntra, at five photos each and a higher resolution, are ₹3,000. The smallest pack is ₹100 for 10 credits, which is enough to photograph 10 designs at 1K and judge the output on your own fabric before spending more. The full rate card, including video, is on the pricing page, and the arithmetic against studio rate cards is in our guide to product photography cost in India.

Questions saree sellers ask