StableBrowse

StableBrowse is a physical AI data infrastructure service that builds multimodal physical-world datasets for robots and embodied agents. It addresses the shortage of real-world examples of hands using tools, objects changing state, and bodies moving through space by handling the full data layer: capture protocol, sensor synchronization, calibration, depth, hand tracking, hand mesh, camera trajectory, tactile streams, temporal labels, and reviewable delivery.

The service covers multimodal capture with egocentric RGB, stereo depth, IMU, audio, hand pose, tactile gloves, tool state, and environment context, followed by post-processing into aligned outputs including depth, hand tracking, hand mesh, camera trajectory, object state, action boundaries, and sensor QC. Delivery includes synchronized timestamps, calibration, schema-valid labels, review UI assets, manifests, and provenance. Custom data programs can be built around a specific physical skill, environment, tool, object class, or sensor stack. Packaging details such as plans or trials are not described.