Skip to main content

Human Archive

Category: Robotics / Embodied AI

Multimodal data provider for robotics and world modeling, collecting egocentric and tactile data via gig workers to train physical AI systems Human Archive was founded in 2026. The company is led by Raj Patel. Based in San Francisco, California, United States. Team size: 4-20. Total funding raised: $8.2M. Latest round: Seed. Key investors include Wing Venture Capital; NVP Capital; Y Combinator.

Founded
2026
Headquarters
San Francisco, California, United States
Team size
4-20
Total funding
$8.2M

Value proposition

Collecting and labeling the world's largest aligned multimodal dataset of human physical interaction to train robotics foundation models and general-purpose robots

Products and solutions

HA-Multi (fully aligned multimodal dataset with vision, stereo depth, tactile gloves, body IMUs, wrist cameras); HA-Ego (mono RGB vision and wrist cameras dataset with annotations); Custom hardware rigs (camera caps, tactile gloves, full-body motion capture suits, wrist cameras); Data annotation and QA pipelines; Internal models for policy benchmarking

Unique value

Only company able to synchronize and collect headset RGB-D, force feedback, full-body motion capture, and synchronized chest/wrist camera data at scale across diverse real-world environments (homes, hotels, restaurants, factories, construction, agriculture)

Target customer

Frontier AI labs, robotics foundation model developers, general-purpose robotics companies, and university research labs building embodied AI systems

Industries served

Home services; Hospitality (hotels, restaurants); Manufacturing; Construction; Agriculture; Retail; Industrial

Technology advantage

Custom multi-sensor hardware (7+ products) capturing RGB-D, tactile force, motion capture, and audio synchronously; 1,000+ active headsets deployed; 50,000+ contributor network; 125+ industry partnerships; Proprietary QA, anonymization, and annotation pipelines; Internal model training for data quality validation on robots

How they differentiate

Custom hardware rigs capturing synchronized multi-modal data (vision, depth, tactile, motion capture) — not just video; deployment in diverse real-world environments across India; ability to fine-tune AI models and test on robots to demonstrate data quality

Main competitors

Scale AI (robotics data labeling); Humyn AI; Neocambrian AI; FPV Labs; Objectways

Key partnerships

125+ partnerships across hotels, restaurants, quick commerce platforms, construction sites, and factories in India; Y Combinator (W26 batch); Snabbit (exploratory discussions)

Major milestones

Founded 2026 by Stanford and UC Berkeley dropouts; Accepted into Y Combinator W26 batch; Raised $8.2M seed round (May 2026); Deployed 1,000+ custom headsets; Collected tens of thousands of hours of multimodal data; Launched HA-Multi and HA-Ego datasets; Built 25-person operations team in India

Growth metrics

1,000+ active headsets deployed; 50,000+ contributor network; 125+ industry partnerships; 7+ custom hardware products; 8,000 hours/day data collection capacity

Market positioning

Early-stage entrant in the rapidly growing robotics training data market, differentiated by multi-modal sensor fusion and real-world data collection at scale via gig economy partnerships in India

Geographic focus

India (primary data collection hub); San Francisco (HQ); Asia expansion

About Raj Patel

UC Berkeley dropout; previously a farmer (sold mangoes & planted trees)

Official website: