Parlour-exit camera that scores every cow's walk and body condition daily
TL;DR Gaitmark puts one camera over the lane where cows walk out of the milking parlour. Software scores how each cow walks and how much condition she carries, every single milking, and the farm gets a short list of cows to check. Commercial dairies with 100 or more cows pay a monthly fee per cow.
India produced 239.3 million tonnes of milk in 2023-24[1], and a growing share of it comes from commercial farms with crossbred cows and parlours. Lameness is one of their quiet losses.
Studies put clinical lameness at 20–30% of dairy herds worldwide, and a 2026 study in neighbouring Bangladesh found 15–25%[2]. A lame cow gives 1.5–4.0 litres less milk a day, and in that study yield had fallen 21% by the day of diagnosis[2]. That adds up fast.
Indian data is thin though. A 2026 Haryana welfare study scored lameness but gave no herd rate[3]. So how do most farms spot a lame cow? By eye, and usually late.
What it does, in regulatory terms: monitoring and flagging for farm staff only. Gaitmark does not diagnose disease or recommend treatment. Diagnosis and treatment stay with a registered veterinarian.
The seed wanted autonomous robots patrolling barns. Robots roaming among cattle are costly, they scare the animals and they add very little that a fixed camera can't see, so I reframed it. Every cow already walks the same lane two or three times a day after milking.
Here's how it works - a camera above that lane records each pass and picks out the cow by her ear tag, RFID or coat pattern. Software scores her gait on a standard mobility scale and her body condition from the shape of her back. The farm gets a daily list of cows whose scores have slipped, so the hoof trimmer and vet see them days earlier. Herd trends go into a monthly report for the owner.
A UK system already works this way[4]. Gaitmark adapts it to Indian breeds, buffalo and farm layouts here.
The owner of a commercial dairy with 100 or more cows pays, usually after a vet or nutritionist recommends it. Milk-procurement companies that run farmer-support programmes are a second buyer for clusters of smaller farms.
What does it cost today? Nobody says. There's no public price for camera lameness systems and the UK vendor doesn't publish one[4], so price discovery is a sprint task.
Vision models now run on cheap edge hardware, and peer-reviewed field work shows camera scoring can track lameness. The UK vendor reports a 75% cut in severe lameness on monitored farms, which is a vendor claim worth checking[4]. India's herd of 192.5 million cattle and 109.9 million buffalo[5] is also shifting, slowly, toward larger parlour-milked farms.
There's no public count of Indian dairy farms with 100 or more cows, and no figure for what they spend on lameness. The firm numbers are adjacent - 239.3 million tonnes of milk[1] and a 1.5–4.0 litre daily loss per lame cow from overseas studies[2].
If you want a real number then count parlour farms in two states. Until then treat this as a small, specialist market.
A Belfast firm sells a parlour-exit camera that scores mobility and body condition at every milking[4]. In India, dairy tech firms focus on milk collection and herd data[6], and heat-detection collars cover breeding.
I found no Indian product that scores gait and body condition from video. The gap is a system trained on Indian breeds and buffalo.
| Difficulty | Moderate (47/100) |
| Starting capital | ₹90 L–1.8 Cr |
| Time to launch | 12–18 months |
The hard part is the data. You need thousands of labelled videos of Indian cows and buffalo, scored by trained assessors, before the model is any good. Farm lanes don't help either, because dust, mud, low light and cows walking in pairs all hurt accuracy and none of that shows up in a clean demo. And then the farm has to act on the list, or the scores change nothing.
The software needs no sector licence, as far as I found, but a few rules shape how you sell it.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
AI here only flags cows. It must never diagnose disease or suggest treatment, and scores trained on Holsteins can misjudge buffalo and zebu crosses.
Want to work out the AI layer for Gaitmark? Talk to Yash →
Here's what can go south -
I think this is a solid niche, not a big one. It fits a founder with computer-vision skills and access to large dairies in Punjab, Haryana, Gujarat or Maharashtra. If you don't have a vet partner then skip it.
Prove one number first - the camera's lameness score agrees with a trained assessor on at least 80% of cows on two farms.
Plan for roughly ₹90 L–1.8 Cr, much of it for labelling thousands of videos of Indian cows and buffalo before the model is reliable.
No sector licence for the software, as far as I found. Keep outputs as flags for the farm's vet, get WPC approval for any wireless box and follow the DPDP Act for worker video.
Yes, a UK firm already sells one from the UK. Your edge would be breeds and buffalo common in South Asia, so check local data and vet practice rules first.
The free Gaitmark validation kit has the line-by-line capital breakdown, the effort scoring, the veterinary and data checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for dairy owners and vets, and 5 assumptions to test. Download the kit →
Thinking about building Gaitmark or something close to it? The make-or-break work is building a labelled video set of Indian cows and buffalo that vets actually trust, and that's the kind of thing I help founders plan. 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.