AI concept sketches plus castable CAD files for small-town jewellers
TL;DR Waxwright gives independent jewellers a counter tool that turns a customer's description or photo into a few design concepts. A human CAD designer then takes the chosen one and returns a castable file within 48 hours. Jewellers pay per design, and it works because custom orders are where small shops still beat the chains.
India's gems and jewellery sector is worth about US$85 billion and employs nearly five million people[1], but most of the retail happens in small family shops. Their customers increasingly walk in with a social media screenshot and ask for "this, but lighter". With gold prices high, jewellery buying is weakening while investment demand holds up[1], so a shop that can redesign a piece to a budget keeps the sale.
Most small jewellers can't do that quickly. They sketch by hand, send the idea to a karigar or a CAD freelancer and wait days, and the customer often drifts off in between. I found no public figure for how many custom orders get lost this way.
What it does, in regulatory terms: design software and a CAD service only. Waxwright never sells or holds gold. The jeweller sells the piece and stays responsible for BIS hallmarking, which now covers 392 districts[2].
At the counter, the jeweller types or speaks the brief (a 6-gram ring, a peacock motif, room for a small ruby) or uploads the customer's photo. Within minutes the tool shows four to six rendered concepts with an approximate gold weight for each. The customer picks one and tweaks it. A Waxwright CAD designer then rebuilds it as a file a casting unit can use, checks wall thickness, stone settings and weight, and sends it back with a 3D render the jeweller can share on chat apps.
I reframed the seed. Text-to-3D tools are already free and fast[3], so generation alone is not a business, and in my judgement a raw AI mesh isn't ready to cast in metal. The value is the checked file and the speed.
The jeweller pays and the shopper doesn't. The target is an independent shop or small chain with one to five stores in a tier-2 or tier-3 town. Pricing would be per finished design file plus a small monthly fee for the counter tool. I found no published rate card for freelance jewellery CAD in India, so the price has to come from interviews. The customer pays the jeweller as usual, with GST on jewellery at 3%[1].
Mandatory hallmarking has spread from 256 districts in June 2021 to 392 by August 2026[2], which pushes small jewellers toward cleaner records and software. Image and text-to-3D generation got cheap and fast in the last two years, and one tool claims over 12 million users and exports printable files[3]. Online demand is moving to smaller towns too, with tier-2 and tier-3 cities giving 66% of new D2C orders in FY26[4].
There's no public figure for India's custom-jewellery or jewellery-CAD market. The narrowest useful signal is the count of hallmarking districts (392), each with registered jewellers who make up your buyer pool[2].
On the adjacent side, the whole gems and jewellery sector sits at about US$85 billion (roughly ₹8.2 lakh Cr)[1]. That's context and not your market.
One big jewellery chain runs a public design contest, but its catalogue is its own and it doesn't sell design services to other shops[5]. General text-to-3D tools generate 3D models from text or images for anyone, with no jewellery checks[3].
Plenty of freelance CAD designers serve jewellers over chat apps, but I found no Indian service that combines a counter tool, fixed turnaround and weight estimates for small shops.
| Difficulty | Moderate (37/100) |
| Starting capital | ₹55–95 L |
| Time to launch | 6–9 months |
The software is the easier half, though tuning image models to produce Indian designs that jewellers recognise takes data and taste. The harder half is running a CAD desk that hits a fixed turnaround and getting jewellers, who buy on trust, to change a habit.
Waxwright itself needs no jewellery licence, because it never sells or holds gold. The rules sit with the jewellers you serve, and your files and pitch should respect them.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't trust an AI mesh to go straight to casting, or an AI weight figure as a final price. A CAD designer checks every file, and the jeweller sets the price.
Want to work out the AI layer for Waxwright? Talk to Yash →
I think this suits a founder who knows jewellery retail or CAD and can sell to shop owners face to face. A pure AI founder will struggle, because the moat is the checked file and the turnaround and not the model. Prove first that 20 jewellers in one town will pay per file, every month, for three months.
Plan for roughly ₹55–95 L, split between the design software and a CAD desk that can hit a fixed turnaround.
Not for the design service itself, since you never handle gold. The jeweller stays responsible for BIS hallmarking, and you still need GST and DPDP compliance.
Yes, the tool works for independent jewellers anywhere. Retrain it on local styles and check local hallmarking and design rights before you sell.
The free Waxwright validation kit has the line-by-line capital breakdown, the effort scoring, the hallmarking and data-protection checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for jewellers, and 5 assumptions to test. Download the kit →
Thinking about building Waxwright or something close to it? The make-or-break work is getting jewellers to trust your weight estimates and turnaround, and winning over traditional shop owners is something I enjoy helping with. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.