The cheapest capable model wins. We make sure it does.
Most AI bills are paid sending every request to a frontier model — even the trivial ones. wave Dispatch is a serverless edge router that classifies each request and sends it to the cheapest model that can actually do the job: your local models first, your frontier model (Claude · GPT · Gemini) only when the work genuinely needs it.
Your infra · your keys · your data. We never see your inference — we return a routing decision. You bring the models (local + any frontier, BYOK); we orchestrate at the edge and bill only the decision.
Safe by default, smarter over time. A verify-gate catches weak local answers and escalates instead of shipping them — so local-first is reliable, not risky. We learn the best model for each kind of work, so your routing keeps getting cheaper and faster.