DARK STORES & MICRO-FULFILMENT
Grid robots and shelf-scanning robots on a small floor, served from one Czar placed near the customer, on the store's own network — a Grid when the floor grows.
THE PROBLEM
Where the work is.
A dark store is a warehouse the size of a shop, close to the customer, with robots that need a model larger than they carry and a cloud that is farther than it looks. One Czar on the store's network serves the floor. When the floor grows, more Czars join and it becomes a Grid, with nothing rewritten.
FACILITIES · CZAR GRID

CONSTRAINTS
Why a cloud, or a GPU on every robot, fails here.
A small floor cannot justify a vendor's full cloud contract, or a GPU on every robot.
Item and shelf models learned elsewhere do not know this store's shelves.
Store imagery is the retailer's data.
A site that starts small should not start over when it grows.
DEPLOYMENT
How the Grid deploys.
- 01
Place the pool
Czar nodes in the building's own rack, on the building's own network.
- 02
Connect the fleet
Each vendor's software calls the pool. Safety and motion stay on the machine.
- 03
Run the site's tasks
Authored, approved, and recorded in Agency Studio on the Grid.
REPRESENTATIVE USE CASES
What a site runs on it.
Illustrative capabilities, not accounts of specific deployments.
A small floor, one Czar
Grid robots and shelf-scanning robots on a small floor, served from one Czar placed near the customer, on the store's own network.
A store that learns its own shelves
Item and shelf models tuned on the store's own captures, on the Czar, with nothing leaving the store.
Reading and boundary
READING
A dark store on one Czar, a Grid when it grows
A dark store is a warehouse the size of a shop, close to the customer, with robots that need a model larger than they carry and a cloud that is farther than it looks. One Czar on the store's network serves the floor, and more Czars join as a Grid when the floor grows.
Safety, stops and motion stay on the machine as its vendor built them. Czar Grid serves perception, decisions and adaptation over the site's network; it is not a controller and does not sit inside any real-time or safety loop.