Instawork launches Instacore wearable robot-data rig

Business Insider got the first look at Instacore, a wearable data-collection rig from Instawork, the gig-work marketplace. The unit, as described by Business Insider, uses five cameras mounted on the head, chest, and wrists and connects to a compute backpack sized for an eight-hour shift. The system is designed for Instawork "Pros" to record themselves performing real-world tasks, with footage intended for companies and research labs training robots. Business Insider reports Instawork has raised more than $150 million from investors including Benchmark, Greylock, and Spark Capital, and operates a marketplace of about 10 million Pros. Business Insider reports Instawork declined to name customers but said it works with leading research labs.

Instawork's Robotics Lab describes the hardware as using hardware-time-synchronized cameras: a wide-angle chest camera that captures the full working environment, plus wrist cameras with tracking markers that let computer-vision models estimate the 3D position and angle of each hand, supported by IMUs for pose estimation. The company says it is already collecting hundreds of thousands of hours of real-world task data per month through its Instawork Lens platform, and that Instacore is built to scale that volume and quality.

For robotics training, synchronized head, torso, and wrist feeds increase coverage of hand-object interactions and environmental context, which benefits perception and imitation-learning pipelines. Practitioners gathering similar data typically pair multi-camera streams with time-aligned IMU or pose data and substantial annotation to support supervised or inverse-modeling approaches.

Public reporting places Instacore in a broader industry push for embodied, real-world data as AI moves from purely virtual models to physical agents. Business Insider frames this within larger robotics investments across the sector. For practitioners, access to large, diverse datasets that reflect messy commercial environments can accelerate robustness testing and sim-to-real transfer, but quality, labeling, and privacy controls remain practical bottlenecks.

monitor which labs and companies, if any, are disclosed as partners and what data formats, labels, and access terms are offered. Also watch how Instawork handles consent, anonymization, and worker privacy in deployment, and whether the dataset is offered under reproducible licensing or remains proprietary.