Your first inference call

Send a request through the Allocate Gateway in under five minutes.

Every request on Allocate goes through the Gateway: one endpoint, one key, every model in the catalog, metered from the first token.

1. Get a key

Sign in at allocate.network and open Keys in your dashboard. Keys are scoped to your organization and can be revoked at any time.

2. Send a request

The Gateway speaks the standard chat completions wire, so any client that takes a base URL works unchanged.

curl https://api.allocate.network/v1/chat/completions \
  -H "Authorization: Bearer $ALLOCATE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-3.5-flash",
    "messages": [{ "role": "user", "content": "Say hello." }]
  }'

Or with an SDK:

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.allocate.network/v1",
  apiKey: process.env.ALLOCATE_API_KEY,
});

const completion = await client.chat.completions.create({
  model: "google/gemini-3.5-flash",
  messages: [{ role: "user", content: "Say hello." }],
});

The model field takes any model id from the catalog, or the name of one of your Routes. A Route pins model, fallback, and policy server-side, so your application code never changes when the routing decision does.

3. See what it cost

Open Meter in your dashboard. Every request is metered as it happens: tokens, cost, latency, and the gateway's own overhead, which is reported on every response in the x-allocate-overhead-ms header and held to a 30ms budget.

Next

• Point a Route at your traffic instead of a raw model id.
• Stream responses by setting "stream": true; the wire format is unchanged.
• Set usage alerts before you put real traffic through.