MISSION
Machines that coordinate themselves.
Work anyone can check.
Every fleet of robots and drones today depends on a central server. When the link drops, the fleet stops, machines do the same job twice or not at all, and nobody can prove which machine did what. We are building the system that removes the server, proves the work and learns from it.
THE PROBLEM
One server, one link, one point of failure.
Fleets of robots and drones run through one controller and one radio link. When either fails, every machine waits. Nothing degrades gracefully.
When the work is done, the only proof is the vendor's own log. Nobody else can check it, so the fleet's own data is never trusted or learned from.
THE FIX
Coordination that lives on the machines.
Each machine beacons what it knows to its neighbors. The swarm splits a goal into jobs, does each job once, signs the result, anchors it on chain and learns from the outcome, with no server in the loop.
WHAT WE BELIEVE
Three rules we won't trade away.
01
No single point of failure
If one server or one link can stop the fleet, the fleet is not autonomous. Coordination has to live on the machines themselves.
02
Proved, not hoped
The rule that stops two machines doing the same job is a theorem with a numerical check, not a behavior we trained and hope holds.
03
Work you can check
Every finished job is signed and anchored on a public chain, so a customer can verify it without trusting us or any vendor.
WHERE WE ARE
Live network. Eight months.
No outside money.
8
MONTHS TO LAUNCH
$100
MONTHLY BURN TO GET THERE
202
STARS ON RUMUS
2
FOUNDERS, 100% OWNED
AS OF NOVEMBER 2026 · SOURCE: GITHUB →
SUM Chain, SNIP and OmniNode are live, RUMUS is open source, and an early SwUM prototype runs in Gazebo simulations.
GO DEEPER
