# AI jewellery photoshoot from one phone photo > An AI jewellery photoshoot turns one phone photo of a piece into finished listing and social images without a studio day. In GehnaLab you upload the photo, choose a Look and get 3 options in a few minutes. The AI paints only the scene; the gold, stones and setting stay your own pixels, checked against your upload. Last updated: 2026-10-23 · Canonical: https://gehnalab.com/ai-jewellery-photoshoot ## Key facts - Input: one sharp phone photo on a plain surface, in daylight. - Output: 3 options per Look, in one setting, in a few minutes. - Looks: Marketplace White, Heritage Ivory, Bridal, Daily Wear, Festive, Stone, Velvet. - Price: 5 Looks free, then ₹499 for 10 Looks or ₹1,499 for 30 Looks over 31 days. - Images that fail the piece check are not charged. On-model images are labelled AI look. Photograph a piece on your counter, choose a Look, and get a set of ready images in a few minutes. GehnaLab's AI only paints the scene around your jewellery. The gold, stones and setting in every final image are your own pixels, checked against your upload. ![Kundan and polki choker with a carved emerald drop laid on maroon raw silk beside a brass diya plate and marigold petals](https://gehnalab.com/media/generated/kundan-choker/festive.jpg) *Festive Look made from a museum photo of a kundan choker. (GehnaLab output, Piece unchanged)* ## What an AI jewellery photoshoot actually does A studio shoot for jewellery has three costs: the photographer's day, the props and models, and the retouching afterwards. For a shop that adds twenty new pieces a month, that usually means either waiting for a big batch or listing pieces with counter photos. An AI photoshoot replaces the set and the retouching. You still supply the piece, but only as one clear phone photo. Most AI tools take your photo and generate a whole new image from it. That works for clothes, where a slightly different fold is harmless. It does not work for jewellery. A regenerated necklace can gain or lose a stone, the filigree softens, kundan foil turns flat, and 22k gold drifts towards brass. A customer who orders that image and receives your real piece will notice. GehnaLab works the other way round. The piece is cut out of your photo first and locked. The AI then paints only the surroundings: the silk, the brass plate, the window light, the soft shadow. Your original pixels are placed back on top, and every image is compared with your upload before you see it. ## From counter photo to finished set ![Phone photo of two gold rings, one with a red and green stone and one with a pearl, lying on patterned cloth](https://gehnalab.com/media/phone/rings-on-cloth.jpg) *Before: the phone photo.* ![The same two gold rings on maroon silk with a brass diya plate and marigolds behind them](https://gehnalab.com/media/generated/rings-pair/festive.jpg) *After: the Festive Look. (GehnaLab output, Piece unchanged)* 1. Photograph the piece on a plain surface in daylight, filling most of the frame. A window is enough; a ring light is not needed. 2. Upload it in the studio and tell us what it is: necklace, earrings, ring, bangle, pendant or set. 3. Pick a Look. Marketplace White for Amazon and Flipkart, Heritage Ivory for your website, Bridal for wedding season, Daily Wear for lightweight gold, Festive for Diwali and Akshaya Tritiya. 4. Pick and download. Each Look returns 3 options in that setting; choose the best one and download it. Run another Look on the same photo for a different setting. A job usually finishes in under two minutes. You can close the page and come back; the images wait in your account. ## The pixel check, and why it matters After the scene is painted, GehnaLab compares the piece region of the final image with your upload pixel by pixel. It also checks the edge where piece meets scene for halos and colour seams, and checks that the piece was not blurred or resized in a way that loses detail. Images that pass carry a small Piece unchanged mark. If anything on the piece moved, the image does not get the mark. We still show it, labelled as an AI look, because sometimes the mood is useful for a story post. You always know which is which. Images that fail our checks are not charged. This matters most for the pieces where a redraw is easiest to spot: [kundan and polki](https://gehnalab.com/kundan-polki-jewellery-photography) with their uneven uncut stones, [temple jewellery](https://gehnalab.com/temple-jewellery-photography) with fine nakashi carving, and anything with a hallmark or a stone count that the customer will check. ## Which Looks to use where | Look | Best for | Where it goes | | --- | --- | --- | | Marketplace White | Every piece | Amazon.in, Flipkart and Meesho main images | | Heritage Ivory | Kundan, temple, antique finish | Your website, catalogue PDF | | Bridal | Polki, bridal sets, chokers | Wedding season ads, lookbooks | | Daily Wear | Lightweight gold, silver, chains | Instagram, WhatsApp Business catalogue | | Festive | Coins, bangles, gifting pieces | Diwali, Dhanteras, Akshaya Tritiya posts | If you only sell on marketplaces, start with the [white background guide](https://gehnalab.com/jewellery-photo-white-background). If you sell on Instagram and WhatsApp, the [Instagram guide](https://gehnalab.com/jewellery-photos-for-instagram) covers sizes and what to post. ## What it costs Every new account gets 5 Looks free, with no card. After that, a Look pack is ₹499 for 10 Looks, which works out to about ₹50 a Look. Shops that list every week can take the Jeweller plan at ₹1,499 a month for 30 Looks. A plain white-background cleanup costs about ₹12 an image. All prices include taxes; the [pricing page](https://gehnalab.com/pricing) has the current numbers. For comparison, a studio day in a metro city usually means a photographer, a retoucher and, for on-model shots, a model and a stylist. AI does not replace a campaign shoot for your flagship collection. It replaces the everyday listing work that piles up between those shoots. ## What it cannot do - It cannot fix a blurred or badly lit photo. Because we never redraw the piece, a soft photo gives a soft piece. Use daylight and hold the phone steady. - On-model images are harder. When a piece is placed on a neck or ear, the AI sometimes has to redraw part of it to make it sit naturally. Those images are labelled AI look. See the [model photoshoot guide](https://gehnalab.com/jewellery-model-photoshoot) for how we handle this. - It will not invent a matching pair. If you photograph one earring, you get one earring. Want to see real results before signing in? The [demo on the home page](https://gehnalab.com/#demo) shows outputs from six pieces, including the ones that did not pass the check. ## Questions jewellers ask ### Will the AI change my jewellery design? No. The piece is cut out and locked before the scene is made, and your original pixels are placed back on top. Each image is checked against your upload; only images where the piece matches get the Piece unchanged mark. ### Do I need a DSLR or a lightbox? No. A recent phone camera in daylight is enough. Put the piece on a plain cloth or paper, fill most of the frame, and avoid coloured light and hands in the shot. ### How many images do I get from one Look? Five: a square white hero for marketplaces, two 4:5 scenes, a festive variant and a 9:16 story for Instagram and WhatsApp status. ### Kya AI photoshoot ke liye studio jaana padega? Nahi. Apne counter par phone se ek saaf photo lijiye, upload kijiye aur Look chuniye. Images kuch minute mein ready ho jaati hain. ### Who owns the images? You do. Use them on your store, marketplaces, social media and print without crediting GehnaLab. ### Are my designs used to train AI models? No. Uploads are used only to make your images and are not shown to other customers or used for training.