From key to first token

One base URL. One key. A complete request you can run in about a minute, without creating an account.

OneRouter is an OpenAI-compatible gateway. You change the base URL and the key. Nothing else in your code moves.

Every value below is literal except the key itself.

Start without a deposit#

The Open Tier serves a small daily allowance at a $0 charge, with no deposit and no money hold. It is the fastest way to prove your client is wired correctly before any money is involved.

The example on this page uses a paid model, so it also exercises the billing path.

Your first request, end to end#

Set your key — one bearer key. Accountless keys work.
Set the base URL — keep /v1; your client appends the endpoint.
Send a request — choose a client, copy, then stream.

curl https://onerouter.network/v1/chat/completions \
  -H "Authorization: Bearer $ONEROUTER_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "meta-llama/llama-3.3-70b-instruct",
    "messages": [{"role":"user","content":"ping"}],
    "stream": true
  }'

from openai import OpenAI

client = OpenAI(
    base_url="https://onerouter.network/v1",
    api_key="or-live-...",
)

stream = client.chat.completions.create(
    model="meta-llama/llama-3.3-70b-instruct",
    messages=[{"role": "user", "content": "ping"}],
    stream=True,
)
for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="")

import OpenAI from "openai";

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

const stream = await client.chat.completions.create({
  model: "meta-llama/llama-3.3-70b-instruct",
  messages: [{ role: "user", content: "ping" }],
  stream: true,
});

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? "");
}

client := openai.NewClient(
    option.WithBaseURL("https://onerouter.network/v1"),
    option.WithAPIKey(os.Getenv("ONEROUTER_KEY")),
)

stream := client.Chat.Completions.NewStreaming(ctx,
    openai.ChatCompletionNewParams{
        Model: "meta-llama/llama-3.3-70b-instruct",
        Messages: []openai.ChatCompletionMessageParamUnion{
            openai.UserMessage("ping"),
        },
    })

The base URL is https://onerouter.network/v1

Keep the /v1. Do not append /chat/completions to a client's base-URL field — the client adds the endpoint itself. Trimming the /v1, or adding the endpoint by hand, is the most common setup failure by a wide margin.

Read the receipt#

Every response carries the facts you need to reconcile it. Nothing about the body changes: it is the upstream's response, byte for byte.

Header	Meaning
`x-onerouter-model`	The model that actually answered
`x-onerouter-provider`	The host that served it
`x-onerouter-cost-usd`	Exact charge for this request
`x-onerouter-balance-usd`	Balance after it settled
`x-onerouter-ttft-ms`	Time to first token, measured at the edge
`x-onerouter-receipt`	Stable id for this route decision

A stream reports its charge as usage.cost in the final usage chunk instead — send stream_options.include_usage. See Streaming.

Per client#

Client setup has literal configuration for Claude Code, Codex, Cline, Continue, Zed, Aider, LiteLLM, SillyTavern and any other OpenAI-compatible client.

When it fails#

Every error uses OpenAI's error envelope and carries a stable code, and every message ends in a URL that resolves to an anchor on the errors page.