The 3D map robots need.
Too important for one company to own.
Delivery bots, humanoid robots, AR glasses. They all need centimeter-accurate 3D data. Others proved people will capture it for free. We think they should own what they make and earn for the work.
50,000+ people proved it works
Capturing 3D for the magic of it
Already proven
People will capture for free.They should earn for it.
Scaniverse proved the model works. 50,000 people captured 3D scans with zero financial incentive. Just the magic of seeing their world in 3D.
But those 50,000 people? They own nothing. They earned nothing. The platform kept it all.
Capture for free, own nothing
Their model
Capture for free, own everything, earn ecko
Our model
March 2025
Niantic sold Pokemon GO — the biggest mobile game in history. They kept Scaniverse.
They sold the game
Kept the map
We agree the map is what matters. We disagree on who should own it. The people who build it should.
Why this matters
AI is running out of training data. Robots have none.
The internet has been scraped dry. Every book, article, Reddit post. Consumed. But AI knows almost nothing about physical space. And the robots coming to navigate the real world? They have no 3D map.
The 3D map of the world is the training data.
It's not just a map for navigation. It's the dataset AI needs to understand physical reality. The dataset robots need to learn how to move through the world. And right now, it doesn't exist.
What exists today
2D maps show addresses
Streets and locations. No physical detail.
What mission planning needs
3D spatial understanding
Know the layout before you send the robot.
Mission planning before dispatch
Before you send a robot across town: Will it fit through the entrance? Are there stairs? What route through the building? You need to know before dispatch, not when it arrives.
AI training hits a wall
LLMs scraped text and images. But they have almost no 3D spatial data. That's why they fail at physical reasoning. The data doesn't exist.
Simulation needs real places
NVIDIA says simulation will replace real training "a million to one." But you need actual 3D environments to simulate in.
AI runs on training data. 3D is the kind that doesn't exist yet.
Scale.ai doesn't build AI. They don't make robots. They provide training data — and the entire AI industry runs through them. 3D spatial data is the category that barely exists yet.
The real cost of "free"
They take your labor.Then your compute.
When Scaniverse processes on your phone, you provide the capture AND the processing power. They get both for free. And the quality? Capped at what a phone can handle.
With Stratis: Community-powered processing
Like Bitcoin mining, but useful work. You own it. The work earns ecko.
Why the map matters
GPS for everywhere.Including indoors.
GPS tells you where you are within a few meters. But it doesn't work indoors, can't tell floors apart, and isn't accurate enough for robots.
The 3D map is the VPS infrastructure.
Every robot, every pair of AR glasses, every autonomous system queries the map. The people who captured it own it.
Quality matters
Real GPUs. Real quality.
Your phone balances processing with battery, heat, and other apps. Dedicated GPUs have one job: maximum quality.
Ownership
You build it. You own it.
Google, Niantic, NVIDIA ask you to contribute. None share ownership. With Stratis, your captures are yours. On-chain. Forever.
How it works
Capture. Claim. Earn.
Your phone is a $50,000+ 3D scanner. Cameras, GPS, LiDAR, accelerometers. We've unlocked that sensor array for 3D capture. Completely free. You own the claim.
You capture
Phone sensors
Network processes
GPU workers
Claim recorded
On-chain
Companies spend ecko
API access
You earn ecko
For capture work
$50k+ scanner in your pocket
Walk and capture
Cameras, LiDAR, GPS, motion, altitude. We use them all.
The result
3D data robots can use
Millions of points. Exact dimensions and layouts.
Who will pay for this data?
Any company that needs to plan missions, assess feasibility, or simulate deployments before sending robots. You can't dispatch a fleet to locations you know nothing about.
Every type of robot needs this
Different robots. Same problem. No pre-mission intel exists.
Factory humanoids
Pre-map facility layouts
Delivery robots
Plan routes before dispatch
Service robots
Know building layouts in advance
Warehouse systems
Simulate before deploying
The digital twin
A complete 3D copy of reality
The economics
Imagine if the community owned Scale.ai
Scale.ai just labels training data, and the entire AI industry depends on it. Everything they produce belongs to the company and its shareholders — none of it to the people who made the data. We're building the version where the people who create the data own it.
On every other platform, you create and they own.
On-chain attribution
Every capture is cryptographically attributed to you. Permanent. Verifiable. Yours.
Full portability
Export to Blender, Unity, Unreal. Any format, any time. Your data, your choice.
Earned through work
Capturing earns ecko — the same token developers spend for API access and processing. Earned by work, spent on use.
Niantic kept this. Not Pokemon GO.
When Niantic sold Pokemon GO — the biggest mobile game in history — they kept one thing: the 3D mapping technology. The map mattered more than the game.
By 2030, millions of robots will navigate the physical world. The map will be crowdsourced. The only question is whether the people who build it will own what they create.
We're building the version where they do.
The map assembles one capture at a time. Every capture lands already placed and at true scale — a tile of the world, owned by the person who walked it.
The map every robot will need, built by the people who live on it.
Start capturing today.
It's completely free.
Download the app. Walk around your neighborhood. Own a piece of the most important map that will ever be built. No fees. No catch. Just capture.