Mecka recruits workers for robotics

Mecka is turning robotics data collection into a labor marketplace, which matters because the hard part in robotics is not labeling pixels after the fact, but getting real people to perform useful physical tasks in the first place. Instead of asking contractors to annotate existing sensor logs like Scale did for autonomous driving, Mecka uses Atlas Capture to recruit workers into a consumer app, pay them by the hour, and capture first person video plus hand, force, and motion signals while they do chores or job specific work.1 2 3

• Scale’s original workflow started with data that already existed, then paid a distributed workforce to mark objects in LiDAR and image frames. Mecka moves one step upstream. It has to source the behavior itself, so recruiting and task design become part of the product, not just quality control and annotation software.2 3
• That is why Mecka mixes a consumer style app with a B2B collection arm. The app is good for broad household tasks like folding laundry or making coffee. Partnerships with bakeries, repair shops, and warehouses are how it gets higher value workflows that are harder for a lab to stage on its own.1 3
• The closest comparable is XDOF, another specialist selling robotics training data and teleoperation services. But the broader pressure comes from labor networks like Scale, Mercor, and micro1 moving into robotics, because they already know how to recruit workers, route tasks, review output, and sell packaged data to frontier AI customers.1 2 4

The next phase is that data vendors will not stop at selling recordings. As robotics labs push toward real deployments, the winning companies are likely to bundle worker recruitment, capture hardware, dataset delivery, teleoperation, and eventually robot operations, so that customers buy a full pipeline for teaching machines physical work rather than a pile of raw clips.1 4