Orchestrating…
Scattered-Systems builds Eryon — a topological computational substrate that turns your own devices into a self-organizing personal cloud — and Proton, the portal that makes it yours.
Eryon produces capabilities; Proton makes them addressable. It is the portal where your own devices become a personal cloud you can compose, customize, and own.
Aggregate your devices — a phone, a laptop, a Raspberry Pi — into one cloud, with no architectural seam between a single device and many.
Personal clouds join into community networks, so you can reach past your own hardware without an architectural break.
Eryon coordinates, balances, and heals from the geometry up — no data center to run, no central scheduler to babysit.
See your cluster’s topology and consistency in real time, read straight off the substrate rather than guessed at.
Shape your interface from composable, generative building blocks — pin what works, rebuild what doesn’t.
Your cluster is bound to you by its shape, its data, and your key. Flat data has no shape to hash; yours does.
Eryon treats computation as a topological–algebraic–harmonic object rather than a flat data structure. Its unit of compute — a “plant” — is a tiny universal machine whose state space is a triad on a generalized Tonnetz.
The mesh in the hero is the generalized Tonnetz — Eryon’s configuration space. Each plant’s headspace is one of its 48 rooted triads; the neo-Riemannian group walks between adjacent triads, and those moves are the substrate’s coordinates.
Conventional systems treat data as flat, computation as isolated state-transitions, and coordination as something a scheduler imposes from above. Eryon recovers the structure that approach throws away — the relations between computational positions and the topology of the computational space itself.
The 48 rooted triads of the generalized Tonnetz form a configuration space, and the neo-Riemannian group — P, L, R, and a fifth-shift — acts on it simply transitively. One set of symmetries is therefore the substrate’s geometry, its routing topology, and its addressing scheme at once. Transformations are coordinates.
Nothing is centrally scheduled. Coordination emerges from topology-aware gossip, load balances as a diffusion process, and consistency is read off with sheaf cohomology. The substrate works at any scale — from a single triad upward.
Compute lives on a manifold of tiny universal machines, sliced from one twelve-tone alphabet by the simplicial structure of the Tonnetz.
The neo-Riemannian group acts simply transitively on the 48 triads — so every position is reachable by a unique transformation, and transformations are coordinates.
No central scheduler. Coordination, load-balancing, and regime changes emerge from local dynamics rather than authored control.
Agreement is a measured quantity: sheaf cohomology says what the cluster agrees on and where a global view fails to glue.
Eryon is the foundation. Each project above it is a higher-order service the substrate’s design naturally pulls into existence — and each can stand on its own merits.
A self-resolving distributed computational fabric built on topology, the neo-Riemannian group, and sheaves.
Where you reach, compose, and inhabit your substrate-based personal cloud.
Automated context management — context as a section of the substrate’s sheaf, not a flat key-value store.
Durable state persisted by simplicial position and content-addressed, kept as thin as possible.
The hybrid-consensus fabric where independent clusters compose and authenticate.
Compute service over the composed, multi-cluster fabric.
A decentralized social space and gig economy over the substrate’s identity and consensus primitives.
A framework for composable, dynamic visual experiences built over the substrate.
Scattered-Systems is a computational research and engineering studio built on a single conviction: the conventional architecture of distributed computing is a partial expression of what computation can be. A foundation that takes the topological, algebraic, and harmonic structure of computation seriously produces qualitatively better systems.
Our mission is to democratize distributed computing — to give individuals, small teams, and resource-constrained users access to computational fabrics that conventional infrastructure reserves for enterprise scale. A phone, a laptop, or a Raspberry Pi should be enough to take part.
Founded by Joe McCain III. We treat mathematical rigor as a discipline, not a flourish — and we would rather understate a claim and surprise you than overstate one and let you down.
Eryon and Proton are early. If you want a personal cloud that scales from one device up — or you build serious distributed systems — let’s talk.