The gig economy in India is booming. Millions of workers now zip around on scooters, deliver hot meals, and show up at your doorstep to clean, cook, and fix things around the house. It’s become a massive labor pool, one that’s attracted attention from investors and tech giants alike. But a Silicon Valley startup called Human Archive sees something else in all that activity: free, real-world training data for robots.
Here’s the pitch. Robotics companies and AI labs are racing to build machines that can do physical tasks in the real world. The problem? They don’t have enough video of humans actually doing those tasks. Not just any video, but egocentric footage, the kind you get from a first-person perspective. Human Archive figured out a clever workaround: partner with home services companies in India, have workers wear specially designed caps with cameras, and collect that footage while they work.
The startup recently locked in $8.2 million in funding from Wing Venture Capital, Y Combinator, and a roster of angels from companies like OpenAI, Nvidia, and Google. Not bad for a company that basically started as a scrappy college project.
The Founders and the Vision
Human Archive was cooked up by four students: three from UC Berkeley and one from Stanford. Samay Maini, Rushil Agarwal, Shloke Patel, and Raj Patel (the cousins, with Raj serving as CEO) all came from research backgrounds in robotics and hardware. They spotted a bottleneck in the AI industry and decided to exploit it.
“Robotics labs and frontier AI companies are racing to build machines that can perform physical tasks in the real world, but they face a critical bottleneck — a shortage of high-quality, real-world training data showing humans doing everyday work,” Patel explained. That’s the core bet: Indian gig workers, with their massive and growing presence, represent an untapped source of exactly that data.
Right now, Human Archive has more than 1,000 active headsets deployed across multiple locations in India, working with partners in home services, hotel, and restaurant sectors. They’re not naming the specific companies, but they’re clearly finding willing participants.
The Pushback
Here’s where things get interesting. Human Archive approached several major Indian home services companies for partnerships, including the likes of Urban Company and Pronto. They got rejected. Hard.
The rejection became public last weekend when Indian outlet Entrackr reported that Pronto was actively seeking partnerships to collect worker data for robotics training and that Snabbit had held early discussions with Human Archive before things fell apart.
Urban Company CEO Abhiraj Singh Bhal didn’t mince words on X, saying his company would not engage in such arrangements. Patel fired back that Urban Company would soon be forced to reconsider or risk losing relevance to customer churn. Co-founder Rushil Agarwal was even blunter, posting that Pronto founder Anjali Sardana had laughed at him and called him “stupid” when he raised the idea of a data partnership. Pronto acknowledged the conversations but said it chose not to move forward.
It’s a messy situation, and it raises real questions about how worker data is being handled in India’s rapidly expanding gig economy.
More Than Just Video
What makes Human Archive stand out from other egocentric data collection efforts is the breadth of what they’re capturing. They’re not just collecting video. The startup is using and developing additional devices like tactile gloves, a full-body motion capture suit, and wrist cameras to capture motion and tactile force, all synchronized with RGB-D data.
“To capture data, we started with iPhones; then we built our own custom rigs and caps. Now we have more than seven different hardware products that we use interchangeably across different modalities,” Patel said.
This multisensor approach is their differentiator. Zach DeWitt, a partner at Wing VC, put it bluntly: “No one else in the world has been able to synchronize and collect headset RGB-D, force feedback, full-body motion capture, and synchronized chest and wrist camera data at scale.”
The company now has more than 50 different devices deployed and is working on custom hardware that captures different data points simultaneously. They’re also developing ways to fine-tune AI models with their own data and test them on robots to evaluate task effectiveness.
The Economics and the Ethics
Human Archive pays workers a base rate of $1 per hour for participating in data collection. That’s notably less than what other companies reportedly pay, which ranges from roughly $2.63 to $4.20 per hour according to one Economic Times report. Patel acknowledged competitors pay more but argues Human Archive’s on-the-ground presence in India allows it to keep compensation lower while offering flexible earning opportunities.
For customers, Human Archive has a clever workaround with some partners. When a worker arrives at a home, consumers are offered a choice through the app: pay a discounted price in exchange for consenting to data collection, or pay the full price for an unrecorded visit. Patel says customers have been happy to opt for the former, since video recordings can help resolve disputes about service quality.
But there are legitimate privacy concerns. It’s not entirely clear what information Human Archive gives workers about how their footage is used. The company says its contracts are compliant with India’s Digital Personal Data Protection Act, that it displays privacy policy notices and consent information, and that all data is anonymized with faces blurred. Still, India’s Ministry of Electronics and Information Technology is reportedly looking into the consent mechanisms and data-collection practices of startups in this space.
What’s Next
Human Archive is already expanding beyond India into Southeast Asia and the U.S. They’re building a platform that would let anyone participate in data collection and earn money. For the U.S. market, they want to offer services like cleaning or cooking in exchange for data collection by participating workers, though these programs are still in early pilot stages.
The larger picture is that multiple well-funded startups are racing to build physical AI, and they all need massive amounts of training data showing humans at work. Human Archive is positioning itself as a key player in serving that demand. Whether its approach can scale will depend on the partnerships it strikes and the uniqueness and volume of data it can collect.
The company faced rejection from some major players, but it’s found ways to work around that by teaming up with smaller startups and directly incentivizing workers and customers. It’s a scrappy approach, and honestly, the ethical questions around consent and worker compensation deserve way more scrutiny than they’re getting.
One thing’s for sure: the race to teach robots how to do household chores is getting weird, and it’s happening now.
