Now accepting early access requests

Human Demonstrations
for the Next Generation
of Robots

Collect, structure, and scale real-world human actions for robotics and embodied AI training.

Who are you?

or scroll to see how it works ↓

Human Task

Real-world actions

Video Upload

Phone-mounted camera

AI Processing

Quality + labeling

Robot Dataset

Structured training data

The Problem

Robots Need
Human Data

Modern robotics and humanoid AI systems require large amounts of real-world human demonstrations to learn from. Unlike language models trained on text from the internet, robots need physical action data that simply doesn't exist at scale.

Collecting this data today is expensive, slow, and impossible to scale. Research labs send employees to perform tasks repeatedly. Companies spend millions on specialized hardware. Progress is bottlenecked not by algorithms, but by data.

$50B+

Robotics market by 2030

Goldman Sachs, 2023

10M+

Demonstrations needed per robot

Google DeepMind estimates

100x

Faster data collection at scale

Internal projection

RoboData solves this. We connect thousands of everyday contributors with robotics companies — creating a scalable, cost-effective pipeline for real-world demonstration data.

The Process

How It Works

From real-world task to structured robotics dataset in four steps.

Step 01

Record

Anyone can do it.

A contributor clips a phone onto a head or chest mount and performs a normal household task — vacuuming, folding laundry, cooking. The camera captures a continuous first-person perspective, the same viewpoint a robot would have.

Equipment:Any smartphone or GoPro
Task length:5 – 30 minutes
Perspective:First-person (egocentric)
1 of 4
📱

Phone mounted

Head or chest position

REC
🧹 Vacuuming
👕 Laundry
🍳 Cooking
🍽️ Dishes
Contributors in Action

Real People. Real Tasks. Real Data.

Contributors wear a phone-mounted camera and record everyday household tasks — exactly as a robot would need to learn them.

Contributor watering plants with head-mounted GoPro camera

Watering Plants

Living room — over-shoulder POV

REC
Contributor vacuuming living room with head-mounted phone camera

Vacuuming

Living room floor

REC
Contributor loading dishwasher with head-mounted phone camera

Loading Dishwasher

Kitchen — over-shoulder POV

REC
Contributor folding laundry with head-mounted phone camera

Folding Laundry

Clothing organization

REC
Contributor cleaning bathroom with head-mounted camera — dual view

Bathroom Cleaning

First-person + operator view

REC
📱

Head-mounted camera

Any smartphone or GoPro

👁️

First-person perspective

Captures exactly what the robot will see

🤖

Robot-ready format

Structured for imitation learning pipelines

For Contributors

Your Chores Are
Worth Money.

You vacuum, fold laundry, and clean the kitchen anyway. Clip a phone mount on your head and get paid every time you do — no special skills, no schedule, no boss.

Join Contributor Waitlist

Paid Data Collection

Earn $5–$20 per completed task, paid weekly.

Flexible Schedule

Work when you want, from wherever you are.

Performance Bonuses

Top contributors earn bonuses for high-quality footage.

Work From Home

Tasks are everyday household activities you already do.

No Special Equipment

Any smartphone or GoPro — we guide the setup.

For Contributors

You Already Do
These Chores.

Strap on a phone mount and get paid while you do them.

Task

Time

You Earn

🧹Vacuum the living room

15 min

$14

👕Fold a load of laundry

12 min

$12

🍽️Load the dishwasher

8 min

$8

🛁Clean the bathroom

20 min

$18

🥞Prepare breakfast

15 min

$15

👔Organize your closet

25 min

$20

Do 3–4 tasks a day

$150 – $400/ month

For chores you were already going to do anyway.

📱

Clip on the mount

Any smartphone or GoPro

🧹

Do your chores

Like you normally would

💸

Get paid

Weekly direct deposit

No experience neededNo bossNo commuteWork from homeFlexible hoursCancel anytime
Example Tasks

Tasks You Can Record

Things you already do at home — now with a phone on your head and $5–$20 in your pocket every time you finish one.

👕

Fold Laundry

$12

per task

Easy10–15 min
clothesfoldingsorting
🍳

Clean Kitchen

$18

per task

Medium15–25 min
surfacesdishesappliances
🍽️

Load Dishwasher

$8

per task

Easy5–10 min
dishesloadingorganizing
🧹

Vacuum Living Room

$14

per task

Easy15–20 min
cleaningfloorsfurniture
👔

Organize Closet

$20

per task

Medium20–30 min
hangingfoldingsorting
🥞

Prepare Breakfast

$15

per task

Medium10–20 min
cookingpouringplating

+ Opening doors, making beds, watering plants, and 50+ more tasks coming soon

For Robotics Companies

High-Quality Demo Data
at Scale

Stop spending months setting up internal data collection infrastructure. Access a growing library of annotated human demonstrations — or commission exactly the data your models need.

Request Early Access

Data formats supported

HDF5JSONMP4ROS2LeRobotRLDS

Curated Human Demonstrations

High-quality, first-person video demonstrations from vetted contributors performing real household tasks.

Task-Level Filtering

Filter by task type, environment, duration, contributor demographics, and more.

Structured Metadata

Every dataset ships with action labels, object tracks, timestamps, and quality scores in standard formats.

Quality Scoring

AI-verified quality metrics so you only pay for data that meets your training standards.

Large-Scale Campaigns

Need 10,000 demonstrations of a specific task? We can run targeted collection campaigns on demand.

Our Vision

Building the Data Layer
for Physical AI

We believe the future of robotics will be powered by large-scale human demonstration datasets. RoboData aims to become the infrastructure layer connecting millions of contributorswith the world's leading robotics companies.

Today

Building the contributor network & data pipeline

2026

10,000 contributors, 1M+ demonstrations

2027

50+ robotics company partnerships

2028

The data layer for physical AI

Investors

We're Building the
Data Layer for Physical AI.

Robotics is having its ImageNet moment. The companies that win won't just build better robots — they'll have access to better training data. We're raising a seed round to own that layer.

Massive, Urgent Market

Humanoid robots will require billions of training demonstrations. The bottleneck isn't hardware or algorithms — it's data. We're building the supply chain before demand fully arrives.

Data Flywheel

More contributors → more task variety → better datasets → more robotics buyers → higher payouts → more contributors. A two-sided marketplace with compounding network effects.

Scalable Labor Model

Unlike lab-based data collection, our contributors work from home using hardware they already own. Marginal cost per demonstration falls as the network grows.

Infrastructure Play

We're not building one robot. We're building the data layer every robot company will depend on — the AWS of physical AI training data.

Business Model

Marketplace take rate20–30% on dataset sales
Enterprise campaignsCustom collection at scale
API subscriptionsOngoing data access tiers

Why Now

The humanoid robot wave is here.

Figure, 1X, Apptronik, Agility, Boston Dynamics, and dozens of well-funded startups are all training robots to do household and industrial tasks. Every one of them needs exactly the data we collect. The window to become the default supplier is open now.

Interested in learning more?

We're meeting with investors now. Reach out and we'll share our deck and data room.

Get Early Access

Join the Waitlist

Whether you want to earn or access data, we'd love to hear from you.

Contributors

Start Earning

Record everyday tasks and get paid per completion.

Robotics Companies

Access Demo Data

Get early access to our curated robotics demonstration datasets.