ccproxy already does "intelligent model routing using your logic or ours." Morph's Model Router is a natural "ours" backend.
It classifies a prompt in ~50ms and returns difficulty (easy/medium/hard), ambiguity, and domain (coding/data/design/general), then maps that to a model. One POST, OpenAI-compatible:
const res = await fetch("https://api.morphllm.com/v1/router/multimodel", {
method: "POST",
headers: { Authorization: `Bearer ${process.env.MORPH_API_KEY}`, "Content-Type": "application/json" },
body: JSON.stringify({ input: userInput, allowed_providers: ["anthropic", "openai"], policy: "balanced" }),
});
const { model } = await res.json();/router/multimodel takes allowed_providers and a policy (balanced, cost_efficient, capability_heavy) and returns one model name. ~180ms over HTTP/2. Wiring it into the routing hook lets users route without writing their own classifier.
Docs: https://docs.morphllm.com/sdk/components/router
ccproxy already does "intelligent model routing using your logic or ours." Morph's Model Router is a natural "ours" backend.
It classifies a prompt in ~50ms and returns difficulty (easy/medium/hard), ambiguity, and domain (coding/data/design/general), then maps that to a model. One POST, OpenAI-compatible:
/router/multimodeltakesallowed_providersand apolicy(balanced,cost_efficient,capability_heavy) and returns one model name. ~180ms over HTTP/2. Wiring it into the routing hook lets users route without writing their own classifier.Docs: https://docs.morphllm.com/sdk/components/router