Fresh 2026

Get started

How to send your first LLM request using the Merge Gateway SDK, OpenAI SDK, or Vercel AI SDK.

Merge Gateway is the control plane for production AI. Access all LLMs through a single API, with intelligent routing, cost management, and security built-in. Route to OpenAI, Anthropic, Google, and AWS Bedrock through a single endpoint, with automatic failover and observability.

You'll need a Merge Gateway API key. Get one from the Merge Gateway dashboard.

Merge Gateway SDK

pip install merge-gateway-python

npm install merge-gateway-sdk

from merge_gateway import MergeGateway

client = MergeGateway(api_key="YOUR_API_KEY")

response = client.responses.create(
    model="openai/gpt-5.2",
    input=[
        {"type": "message", "role": "system", "content": "You are a helpful programming tutor. Explain the concepts clearly with practical examples."},
        {"type": "message", "role": "user", "content": "Explain the concept of recursion in programming with a simple set of examples."},
    ],
)

print(response.output[0].content[0].text)

import { MergeGateway } from "merge-gateway-sdk";

const client = new MergeGateway({ apiKey: "YOUR_API_KEY" });

const response = await client.responses.create({
  model: "openai/gpt-5.2",
  input: [
    { type: "message", role: "system", content: "You are a helpful programming tutor. Explain the concepts clearly with practical examples." },
    { type: "message", role: "user", content: "Explain the concept of recursion in programming with a simple set of examples." },
  ],
});

console.log(response.output[0].content[0].text);

If you have a routing policy configured (on a project or as the org default), the model field is optional. The policy picks the provider and model. Either omit model or set it to the sentinel value "default_routing" to explicitly hand off to your policy. To scope a request to a project, pass project_id in the request body.

Swap the model string to route to a different provider. No other code changes needed.

response = client.responses.create(
    model="anthropic/claude-sonnet-4-20250514",
    input=[
        {"type": "message", "role": "system", "content": "You are a helpful programming tutor. Explain the concepts clearly with practical examples."},
        {"type": "message", "role": "user", "content": "Explain the concept of recursion in programming with a simple set of examples."},
    ],
)

response = client.responses.create(
    model="google/gemini-2.0-flash",
    input=[
        {"type": "message", "role": "system", "content": "You are a helpful programming tutor. Explain the concepts clearly with practical examples."},
        {"type": "message", "role": "user", "content": "Explain the concept of recursion in programming with a simple set of examples."},
    ],
)

OpenAI SDK

Already using the OpenAI SDK? Point it at Merge Gateway to get multi-provider routing without changing your application code.

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api-gateway.merge.dev/v1/openai",
)

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "YOUR_API_KEY",
  baseURL: "https://api-gateway.merge.dev/v1/openai",
});

Use the standard chat.completions.create method. No provider prefix needed on the model name.

response = client.chat.completions.create(
    model="gpt-5.2",
    messages=[
        {"role": "system", "content": "You are a helpful programming tutor. Explain the concepts clearly with practical examples."},
        {"role": "user", "content": "Explain the concept of recursion in programming with a simple set of examples."},
    ],
)

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

const response = await client.chat.completions.create({
  model: "gpt-5.2",
  messages: [
    { role: "system", content: "You are a helpful programming tutor. Explain the concepts clearly with practical examples." },
    { role: "user", content: "Explain the concept of recursion in programming with a simple set of examples." },
  ],
});

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

AI SDK (Vercel)

Using the Vercel AI SDK? Use the Merge Gateway provider for full access to routing, tags, and cost tracking, or use the quick-start URL shim for zero-install setup.

Recommended: Merge Gateway provider

npm install merge-gateway-ai-sdk-provider ai

import { createMergeGateway } from "merge-gateway-ai-sdk-provider";

const gateway = createMergeGateway({
  apiKey: "YOUR_API_KEY",
});

Use generateText to send a request. Model names use the provider/model format.

import { generateText } from "ai";

const { text } = await generateText({
  model: gateway("openai/gpt-4o"),
  prompt: "Explain the concept of recursion in programming with a simple set of examples.",
});

console.log(text);

The native provider gives you typed access to Gateway features like tags, vendor routing, and routing metadata via providerOptions.mergeGateway.

Alternative: URL shim (no extra install)

If you already have @ai-sdk/openai installed, point it at Gateway with a base URL change:

import { createOpenAI } from "@ai-sdk/openai";

const gateway = createOpenAI({
  apiKey: "YOUR_API_KEY",
  baseURL: "https://api-gateway.merge.dev/v1/ai-sdk",
});

// All generateText/streamText calls work unchanged

Other SDKs

Gateway is compatible with any SDK that lets you set a custom base URL.

SDK	Base URL
OpenAI	https://api-gateway.merge.dev/v1/openai
Anthropic	https://api-gateway.merge.dev/v1/anthropic
AI SDK (Vercel)	https://api-gateway.merge.dev/v1/ai-sdk
LangChain	https://api-gateway.merge.dev/v1/openai

from anthropic import Anthropic

client = Anthropic(
    api_key="YOUR_API_KEY",
    base_url="https://api-gateway.merge.dev/v1/anthropic",
)

message = client.messages.create(
    model="claude-sonnet-4-20250514",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "Explain the concept of recursion in programming with a simple set of examples."},
    ],
)

print(message.content[0].text)

Next steps

Configure intelligent routing, failover, and cost optimization across providers

Automatically reduce token usage and avoid context window limits

Explore the full Gateway API