Close-up photo and expert checks that tell handloom cloth from powerloom
TL;DR Weftproof checks the finished cloth and ignores the paperwork. A seller or marketplace takes guided close-up photos of a saree's weave and selvedge, a trained textile examiner reviews them with help from an image model, and a dated report comes back. Marketplaces and handloom brands pay per check to catch powerloom substitutes.
Powerloom fabric passed off as handloom is so common that a 2024 study from Assam opens with it. Officials there raided sellers, but without a lab certificate to back the complaint, the seized goods went straight back to the traders[1]. Even experts struggle to tell the two apart by eye[1].
That leaves honest sellers and marketplaces exposed. A brand buying through middlemen can't easily tell whether a batch was swapped, and a marketplace listing thousands of "handloom" sarees has no cheap way to audit them.
Government marks certify producers in small batches, and India Handloom Brand registered 34 products at its December 2024 evaluation meeting[3]. Nothing checks the individual piece once it's out in the market.
The cloth itself carries the evidence. Handwoven fabric tends to show uneven weft picks, temple marks on the selvedge and the odd repaired lump, while powerloom cloth is uniform and often uses untwisted yarn[1]. Weftproof turns those cues into a service.
A guided phone app asks for five close-ups in daylight, at a set distance, of the body, border and selvedge. An image model sorts clear cases from unclear ones. A trained examiner, usually a retired weaver or textile technologist, reviews every report before it goes out, and doubtful pieces get called in for a physical look. The report states a finding and a confidence level and does not pretend to be a legal certificate.
The seed proposed video plus blockchain to prove artisan quality, and I reframed it. A ledger only shows a record wasn't changed and says nothing about whether it was true, and filming every loom is costly. Testing finished cloth works on stock already in the market. One published model reached 94.39% validation accuracy on handloom and powerloom towels from Assam[1], but whether that holds for sarees from other clusters is unproven.
Buyers are handloom brands auditing their suppliers, online marketplaces checking sellers and boutique resellers who want a report for a high-value piece. My estimate is a fee of ₹100–300 per check, with volume plans for marketplaces. State handloom departments could pay for enforcement support later, but that's a slow sale.
More and more handloom sells online, where buyers can't feel the cloth. Handloom marketplaces show Handloom Mark and Silk Mark logos to signal authenticity[4], so trust is already a selling point. And the Assam work shows ordinary smartphone photos, taken at 5–10 cm in natural light, can carry enough detail for a model[1].
There's no public figure for what India spends on handloom authentication. The narrowest evidence is the size of the supply base - Assam alone has 10.9 lakh weaver households, 38.6% of India's total[1], against a few dozen products registered per India Handloom Brand round[3].
That gap shows how little of the market any check reaches today.
India Handloom Brand, which followed the 2006 Handloom Mark scheme, certifies products and producers and not single pieces in the market[2]. Online handloom stores sell cooperative handlooms with certification logos[4]. Academic teams have built classifiers for one fabric type[1].
Another idea in this directory, Loomtrue, tags sarees at the loom. Weftproof checks cloth after it has left the weaver, so honestly the two could work together.
| Difficulty | Moderate (44/100) |
| Starting capital | ₹45–85 L |
| Time to launch | 6–9 months |
The hardest part is building a reference library of verified handloom and powerloom samples for each major saree style, because one model won't cover Kanjeevaram and Chanderi alike. Recruiting examiners who are both expert and impartial comes second. And sellers who fail checks will push back hard.
No licence is needed to give an expert opinion on cloth. The care goes into what your reports claim.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't trust the model on its own for any finding. A trained examiner signs every report, and a published accuracy figure on Assam towels says nothing yet about Kanjeevaram or Chanderi.
Want to work out the AI layer for Weftproof? Talk to Yash →
I'd back this for a founder with textile-testing or handloom-sector experience who can earn the trust of weavers' bodies. If you see it as a pure AI product then it's the wrong idea for you. Start with one saree style, build a verified sample library and prove examiner and model agree on at least 95% of clear cases.
Plan for roughly ₹45–85 L, much of it on building a verified sample library and paying expert examiners.
No licence is needed to give an expert opinion, but you can't call it a statutory certificate or use government handloom marks. Get a lawyer to write the report template.
Yes, wherever handwoven textiles sell at a premium over machine-made copies. You'd need a fresh sample library for local weaves and a check on local consumer law.
The free Weftproof validation kit has the line-by-line capital breakdown, the effort scoring, the claims and data checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for sellers and marketplaces, and 5 assumptions to test. Download the kit →
Thinking about building Weftproof or something close to it? The make-or-break work is a reference library that keeps the error rate low for each saree style, and I'm happy to help you plan how to build it. 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.