How to Use OpenRouter: Access Multiple AI Models with One API

OpenRouter is a unified API gateway that provides access to multiple AI models from different providers like OpenAI, Anthropic, Google, Meta, and many others. Instead of managing multiple API keys and different API formats, you can use OpenRouter to access all these models through a single, standardized interface.

Why Use OpenRouter?

Single API for Multiple Models: Access GPT-4, Claude, Gemini, Llama, and more through one OpenAI-compatible API
No Rate Limits: Bypass the rate limits imposed by individual providers, perfect for high-volume applications
Transparent Pricing: Compare real-time costs across models and choose the best option for your use case
Fallback Support: Automatically switch to alternative models if your primary choice is unavailable

Getting Started

Sign up at openrouter.ai
Create an API key at openrouter.ai/keys
Add credits at openrouter.ai/credits (minimum $5)

Official SDKs

OpenRouter provides official SDKs for easy integration:

TypeScript SDK: npm install @openrouter/sdk - Type-safe access to 300+ models
Python SDK: Full documentation at openrouter.ai/docs/sdks/python
Vercel AI SDK: npm install @openrouter/ai-sdk-provider - For Vercel AI SDK users
OpenAI SDK Compatible: Works with existing OpenAI SDKs by changing the base URL

Using OpenRouter with JavaScript/TypeScript

Using the Official SDK:

import OpenRouter from "@openrouter/sdk";
 
const openrouter = new OpenRouter({
  apiKey: process.env.OPENROUTER_API_KEY,
});
 
const completion = await openrouter.chat.completions.create({
  model: "anthropic/claude-3.5-sonnet",
  messages: [
    { role: "user", content: "Explain quantum computing in simple terms" }
  ],
});
 
console.log(completion.choices[0].message.content);

Using fetch (alternative method):

async function chat(message) {
  const response = await fetch("https://openrouter.ai/api/v1/chat/completions", {
    method: "POST",
    headers: {
      "Authorization": `Bearer ${process.env.OPENROUTER_API_KEY}`,
      "Content-Type": "application/json"
    },
    body: JSON.stringify({
      "model": "anthropic/claude-3.5-sonnet",
      "messages": [{ "role": "user", "content": message }]
    })
  });
 
  const data = await response.json();
  return data.choices[0].message.content;
}

Using OpenRouter with Python

Using the Official SDK:

pip install openrouter

from openrouter import OpenRouter
import os
 
with OpenRouter(api_key=os.getenv("OPENROUTER_API_KEY")) as client:
    response = client.chat.send(
        model="anthropic/claude-3.5-sonnet",
        messages=[
            {"role": "user", "content": "Write a haiku about programming"}
        ]
    )
    print(response.choices[0].message.content)

The official SDK is auto-generated from OpenRouter's OpenAPI specs, provides full type safety with Pydantic validation, and supports streaming and async operations.

Using OpenAI SDK (alternative):

from openai import OpenAI
 
client = OpenAI(
    base_url="https://openrouter.ai/api/v1",
    api_key="your-openrouter-api-key"
)
 
response = client.chat.completions.create(
    model="anthropic/claude-3.5-sonnet",
    messages=[{"role": "user", "content": "Write a haiku about programming"}]
)
 
print(response.choices[0].message.content)

Available Models

OpenRouter provides access to many models. Here are some popular ones:

OpenAI: openai/gpt-4, openai/gpt-3.5-turbo
Anthropic: anthropic/claude-opus-4, anthropic/claude-sonnet-4.5
Google: google/gemini-3-pro-preview, google/gemini-3-flash-preview
Meta: meta-llama/llama-4-scout, meta-llama/llama-4-maverick
Mistral: mistralai/mistral-large, mistralai/mistral-medium

You can find the complete list of available models at https://openrouter.ai/models.

Model Fallbacks

OpenRouter automatically falls back to alternative models if your primary choice is unavailable. This works with both the SDK and raw API:

Using the SDK:

import OpenRouter from "@openrouter/sdk";
 
const openrouter = new OpenRouter({
  apiKey: process.env.OPENROUTER_API_KEY,
});
 
const completion = await openrouter.chat.completions.create({
  models: [
    "anthropic/claude-opus-4",
    "openai/gpt-4",
    "google/gemini-3-pro-preview"
  ],
  route: "fallback",
  messages: [
    { role: "user", content: "Hello!" }
  ],
});

Using fetch:

const response = await fetch("https://openrouter.ai/api/v1/chat/completions", {
  method: "POST",
  headers: {
    "Authorization": `Bearer ${process.env.OPENROUTER_API_KEY}`,
    "Content-Type": "application/json"
  },
  body: JSON.stringify({
    "models": [
      "anthropic/claude-opus-4",
      "openai/gpt-4",
      "google/gemini-3-pro-preview"
    ],
    "route": "fallback",
    "messages": [{ "role": "user", "content": "Hello!" }]
  })
});

If the first model returns an error (rate limits, downtime, etc.), OpenRouter automatically tries the next one.

Best Practices

Store API Keys Securely: Always use environment variables, never hardcode API keys
Monitor Usage: Check your OpenRouter dashboard regularly to track spending
Handle Errors: Implement proper error handling for API failures
Set Timeouts: Add reasonable timeouts to prevent hanging requests
Use Appropriate Models: Choose models based on your use case (speed vs. quality vs. cost)

Pricing

OpenRouter uses a pay-as-you-go model with transparent pricing. Each model has different costs per token. Generally:

Smaller models (like Llama 3 8B) are very cheap
Medium models (like GPT-3.5 Turbo) are moderate
Large models (like GPT-4 or Claude Opus) are more expensive

You can see real-time pricing for each model at https://openrouter.ai/models.

Conclusion

OpenRouter is a practical solution for working with multiple AI models without the overhead of managing separate integrations. The examples above should get you started, but there's much more to explore including custom parameters, fine-tuned models, and advanced routing strategies.

Check out the OpenRouter documentation for more details.