WAVE Mesh is live

Reachability is the problem nobody puts on the slide. You have devices behind home routers, on hotel Wi-Fi, on cellular, behind a venue's locked-down firewall — and you need them to talk to each other and to your infrastructure, reliably, without each operator standing up a VPN or babysitting a tailnet. Most of the time the answer is a pile of third-party overlays, NAT-traversal hacks, and a billing system that has no idea any of it happened.

WAVE Mesh is the opposite bet. It's WAVE's conversion layer for reachability: every WAVE device joins one flat, NAT-regardless, WAVE-owned mesh under a single identity — direct when the net allows, relayed when it doesn't, controlled and metered end to end. Today it's commercially available, priced and billed on the same gateway as everything else WAVE does.

One overlay, one identity

The shape of a mesh decision used to be: pick an overlay vendor, wire its agent onto every device, hand each operator a tailnet to join, and hope the NAT gods are kind. WAVE Mesh collapses that. A device authenticates once with its WAVE identity and it's on the mesh — no external tailnet, no per-operator VPN, no separate account. Reachability becomes a property of being a WAVE device, not a project you run on the side.

From there the coordinator does the work the overlay vendors charge you to think about: it plans routes across whatever interfaces a node actually has, prefers a direct path when the network allows one, falls back to a WAVE relay when it doesn't, and watches heartbeats so a dead path fails over without a human in the loop.

What the mesh carries

Reachability is the floor, not the ceiling. Because the mesh is WAVE-native, the same primitives you already use ride on it:

• Multihoming and diversity-grid reachability — the coordinator sees every interface a node exposes and plans across them, so a node with more than one way out is more reachable, not just differently reachable.
• Stream and device orchestration — start, route, and tear down streams and devices across the mesh as one fan-out operation, not a per-device script.
• Content on the mesh — move recordings and assets between WAVE devices over the same governed, metered path.
• Observability and policy — every route is visible, and routing policies are a first-class, governed control surface with an audit trail.

What it costs

Mesh is metered on four dimensions, billed through the same gateway and the same bill as the rest of WAVE — no separate invoice, no separate vendor:

• Requests routed — $0.0002 / op. The core unit; roughly $6 per device per month in typical use.
• Failover events — $0.02 / event. You pay for reliability when it actually fires.
• Active regions — $2.00 / region·month. Your geographic footprint.
• Routing policies — $1.00 / policy·month. The control surface.

The one honest caveat

We label what's proven and what isn't, so here's the line. The control plane and the M-A → M-G spine are live. The multi-ISP / two-uplink physical braid — true link diversity across two distinct physical uplinks — is implemented in the coordinator's path planner, but it is not yet proven over real dual-uplink hardware. It's on the roadmap, not in the "done" column. Everything else on this page is live; that one is the exception, and we'd rather you know.

Start routing

Point a WAVE device at the coordinator, read the contract, and watch the first ops meter.

See WAVE Mesh Read the docs

✦one identity, every device, NAT regardless