The category
Mesh Native Computing
Software that lives on infrastructure you own. Services meet over a federation of relay stations — no hyperscaler account in the middle, no master node, no single company that can change the terms. The mesh on this site is live: every station, demo and stream here runs on the real fleet.
What it is
Infrastructure you own
Your code runs on your hardware — a mini-PC, a datacenter rack, a tower edge node. When you own the iron, nobody else can change the terms of service.
Local-first
Nodes keep working locally when the connection drops: they record their own events on the spot, and reconnect and resubscribe when the link returns.
Federated, not centralized
No master node. Each organization keeps its own identity — a realm — and decides who peers with whom. One organization's outage doesn't stop the others.
No cloud tax
No per-request fees, no egress charges. The core and SDKs are open source (Apache-2.0), so the code you build on stays yours to read, fork and keep.
What it can do for your business
Mostly it removes things: the middlemen, the single points of failure, the jurisdiction you don't control. Here is what that means by industry.
Telecom operators
Tower-to-tower traffic without a round-trip to a hyperscaler region, and edge nodes that keep working locally when backhaul fails. You sell the rails; your customers keep their own identity.
Datacenter operators
Colocation is commoditizing; sovereign compute is where the growth is. A European operator offers what a US-owned provider can't: no parent company subject to US law. Your racks become the product.
Hosting providers
The VPS you already sell, minus the jurisdiction question. "Sovereign VPS" turns the same hardware into what public-sector and regulated customers are increasingly asked to buy.
Software companies
Deploy to hardware your customer owns. No platform fees on your revenue, and no foreign policy decision can shut down your distribution overnight. Keep your margins, keep your customers.
Enterprises
Your Munich office talks to your Milan warehouse through relays you choose, not a foreign hyperscaler. Where your data travels becomes a design decision you can show an auditor, not a clause in someone else's terms.
Consumers & builders
AI now runs at home: NVIDIA's Personal AI Router turns a household's machines into one inference cluster. Macula links those clusters across the internet — members of a realm you trust can use your cluster, and you theirs, with no cloud account in between. Our first bridge already does this for chat.
See it working
Copy only goes so far. Everything linked below runs on the live fleet — real nodes, real relays, real facts crossing the mesh.
Talk to us about your use case
No pricing page, no self-serve checkout. Tell us what you're building and we'll answer with what's possible — straight from the people building the mesh.