WAVE · Dispatch

Media infrastructure for the agentic internet — routed to the cheapest capable model.

Local-first AI dispatch: classify each request at the edge, run it on your own infra for free when a local model is capable, and escalate to a frontier model only on low confidence. Your infra, your keys, your prompts never leave.

Start 7-day trial (card required) → Talk to sales Create a WAVE account →

try it

curl https://dispatch.wave.online/pricing.json | jq .topology_invariant.per_decision_usd
0.0001

How it routes

Classify at the edge

A request is embedded at the edge (Workers AI, bge-base-en) and classified into a route with a confidence and margin — before any model call happens.

Local first, $0

local_code, local_search, local_summarize, and direct all run on your infra with your keys — no per-token bill to WAVE for the decisions that stay local.

Escalate on low confidence

Only when the classifier is unsure does the request escalate to a frontier model (Claude, GPT, Gemini) — the expensive call is the exception, not the default.

tier 1 · local pooltier 2 · local agenttier 3 · frontier (rare)

  prompt
    │
    ▼   embed @ edge · Workers AI (bge-base-en)
  classify ─▶ {route · confidence · margin}
    │
    ├─ local_code       ┐
    ├─ local_search     │  $0 · runs on
    ├─ local_summarize  ┤  YOUR infra + keys
    ├─ direct           │
    └─ reason           ┘
         │
         └─ low confidence ─▶ escalate to your
            frontier model (Claude · GPT · Gemini)

classifyPOST / {"prompt":"…"} → {route, probability, margin}

authAuthorization: Bearer <license-key> · or x402 pay-per-use

pricingStarter $9/mo · Pro $29/mo · Scale $99/mo · Enterprise custom · x402 $0.001/decision (covers on-chain settlement gas — account billing is $0.0001) · /pricing.json

BYOK — your API keys + data + inference stay on your infra. We only ever return a routing decision.