Quickstart

The shortest route from your key to a working chat completion. About five minutes if you already know curl.

Before you start

You need:

• A gateway base URL — your gateway's address. For local development it's <GatewayUrl/>.
• An API key starting with gw_ — create one in the API keys section. Treat it like any secret (don't commit it, don't paste it into chat).
• A model slug — see the models page. gpt-5.5 is a safe default.

Send a request

The gateway speaks the standard chat/completions shape on /v1. Drop your existing client in front of it; the only change is the baseURL and the key.

cURL

curl -X POST https://api.toygate.store/v1/chat/completions \-H 'Authorization: Bearer gw_<your-key>' \-H 'Content-Type: application/json' \-d '{ "model": "gpt-5.5", "messages": [{ "role": "user", "content": "Say hi" }] }'

TypeScript

import OpenAI from "openai";const client = new OpenAI({apiKey: process.env.GATEWAY_KEY,baseURL: "https://api.toygate.store/v1",});const resp = await client.chat.completions.create({model: "gpt-5.5",messages: [{ role: "user", content: "Say hi" }],});console.log(resp.choices[0].message.content);

Python

from openai import OpenAIclient = OpenAI(  api_key=os.environ["GATEWAY_KEY"],  base_url="https://api.toygate.store/v1",)resp = client.chat.completions.create(  model="gpt-5.5",  messages=[{"role": "user", "content": "Say hi"}],)print(resp.choices[0].message.content)

What you should see

A standard chat-completion response. The model field echoes the slug you asked for, choices[0] contains the assistant turn, and usage reports prompt + completion tokens.

The gateway hid the underlying provider from you — your code never had to care which upstream actually ran the request.

Where to go next

• Key concepts — the model slug, what the gateway hides, the bits worth knowing.
• Authentication — bearer tokens, rotation, the public catalog.
• Streaming — Server-Sent Events, idle timeouts, cancellation.
• Errors — the error envelope and stable codes.
• Rate limits & quotas — request rate, concurrency, AI credit balance, and what to do when a limit trips.

Self-host

ToyGate is open source, and you can run it yourself. The production compose file is docker-compose.prod.yml at the repo root. Once you're up, the ops runbooks under docs/how-to/ops/ cover backup-restore, alerting, certificate renewal, and incident response.