Getting started

Your first API call

ezrouter is a unified API that routes requests to many large language models — claude, gpt, deepseek, glm, kimi, and others — behind a single endpoint and a single API key. The API is compatible with the OpenAI and Anthropic SDKs: point your existing client at our base URL and most code works unchanged.

Configuration

The catalog is gateway-global — every key sees the same models. As of this writing the catalog includes claude (haiku / sonnet / opus 4.x), gpt-5.x, deepseek-v4 (flash / pro), glm-5.1, and kimi-k2.6. Always query /v1/models for the live list rather than hard-coding model IDs.

Pick a surface

ezrouter exposes two parallel API surfaces against the same model catalog. Choose based on what client you already have:

You can mix and match — the same API key works on both surfaces.

Make your first request

A minimal chat completion against claude-sonnet-4-6, via the OpenAI-compatible surface:

curl

curl https://www.ezrouter.dev/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer ${EZROUTER_API_KEY}" \
  -d '{
        "model": "claude-sonnet-4-6",
        "messages": [
          {"role": "system", "content": "You are a helpful assistant."},
          {"role": "user", "content": "Hello!"}
        ]
      }'

Python (OpenAI SDK)

# pip install openai
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["EZROUTER_API_KEY"],
    base_url="https://www.ezrouter.dev/v1",
)

response = client.chat.completions.create(
    model="claude-sonnet-4-6",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello!"},
    ],
)

print(response.choices[0].message.content)

Node.js (OpenAI SDK)

// npm install openai
import OpenAI from "openai";

const openai = new OpenAI({
  baseURL: "https://www.ezrouter.dev/v1",
  apiKey: process.env.EZROUTER_API_KEY,
});

const completion = await openai.chat.completions.create({
  model: "claude-sonnet-4-6",
  messages: [
    { role: "system", content: "You are a helpful assistant." },
    { role: "user", content: "Hello!" },
  ],
});

console.log(completion.choices[0].message.content);

For Anthropic-SDK examples and the Anthropic surface, see the ezrouter API reference.

Notable behaviors

Before integrating in production, read these gateway-wide behaviors — they differ from a stock OpenAI or Anthropic deployment in ways that will trip up naive clients:

• The OpenAI surface always returns SSE. Even without

stream: true, responses arrive as data: {...} chunks terminated by data: [DONE]. Clients calling response.json() directly will fail; use an SSE parser. See POST /v1/chat/completions → Differences from OpenAI.

• No 429 or rate-limit headers. The gateway uses silent

upstream queueing under load instead of returning explicit backpressure. Build clients with timeouts and latency-based backoff, not 429-retry handlers. See Rate limits.

• **Tool-call extraction is reliable on claude-opus-4-7 via the

OpenAI surface today; for other models, prefer the Anthropic surface for agent applications.** See POST /v1/chat/completions → Tool calls.

• Error envelope shape is ezrouter-specific and uses three

distinct error.type taxonomies. See Error codes.

Where to go next