TokenRouter Overview - Intelligent AI Routing Platform

What is TokenRouter?

TokenRouter is an intelligent LLM routing platform that sits between your application and AI providers. It automatically selects the best provider and model for each request based on your optimization preferences, routing rules, and firewall policies.

Key Benefits

• Cost Savings: Automatically route to the most cost-effective provider for each request
• Improved Reliability: Failover to backup providers when your primary choice is unavailable
• Better Performance: Route based on latency, quality, or balanced optimization
• Simplified Integration: One API for multiple providers (OpenAI, Anthropic, Google, Mistral, DeepSeek, Meta)
• Enhanced Security: Built-in firewall rules and content filtering
• Full Observability: Track usage, costs, and performance across all providers

How It Works

TokenRouter wraps the OpenAI Responses API with an intelligent routing layer:

1. Send a Request: Your application sends a standard request to https://api.tokenrouter.io/v1/responses
2. Intelligent Routing: TokenRouter analyzes your request, routing rules, and provider availability
3. Provider Selection: The optimal provider is selected based on your routing mode (cost, quality, latency, or balance)
4. Firewall Enforcement: Your firewall rules are applied to filter or modify the request
5. Response Delivery: A fully OpenAI-compatible response is returned to your application

The router monitors live provider pricing, historical latency, and your account limits to ensure you always get the best trade-off for each request.

Routing Strategies

You can adjust the routing strategy per-request, define deterministic rules in the console, or force a specific provider when compliance requires it.

Auto Routing

Use automatic routing modes for intelligent provider selection:

const response = await client.responses.create({
  model: 'auto:balance',  // Balanced optimization
  input: 'Explain quantum computing'
});

Available modes:

• auto:balance - Balanced trade-off
• auto:cost - Minimize costs
• auto:quality - Maximize quality
• auto:latency - Minimize latency

Specific Model

Request a specific model with a routing mode:

const response = await client.responses.create({
  model: 'gpt-4o:quality',  // Use GPT-4o with quality optimization
  input: 'Write a detailed business plan'
});

Or just specify the model:

const response = await client.responses.create({
  model: 'claude-3-7-sonnet-latest',
  input: 'Analyze this code'
});

Routing Rules

Define rules in the console for deterministic routing:

• Route all requests containing “code” to Anthropic
• Route requests under 100 tokens to cost-optimized providers
• Force specific models for compliance requirements
• Set fallback providers for high availability

Learn more about routing rules →

Supported Providers

TokenRouter supports the following AI providers:

Provider	Models	Special Features
OpenAI	GPT-4o, GPT-4 Turbo, GPT-3.5 Turbo	Function calling, vision, structured outputs
Anthropic	Claude 3.7 Sonnet, Claude 3 Opus, Claude 3 Haiku	Extended context, vision, tool use
Google	Gemini 1.5 Pro, Gemini 1.5 Flash	Multimodal, long context
Mistral	Mistral Large, Mistral Medium	European hosting, function calling
DeepSeek	DeepSeek V3	Cost-effective, code generation
Meta	Llama 4 (special access required)	Open weights, on-premises options

OpenAI Compatibility

TokenRouter is designed as a drop-in replacement for the OpenAI API. If you’re currently using OpenAI, migrating to TokenRouter requires minimal changes:

Before (OpenAI)

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.OPENAI_API_KEY
});

const response = await client.chat.completions.create({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: 'Hello!' }]
});

After (TokenRouter)

import Tokenrouter from 'tokenrouter';

const client = new Tokenrouter({
  apiKey: process.env.TOKENROUTER_API_KEY
});

const response = await client.responses.create({
  model: 'auto:balance',  // Intelligent routing
  input: 'Hello!'
});

The key differences:

• Use TokenRouter API key instead of OpenAI key
• Use responses.create() instead of chat.completions.create()
• Use input parameter instead of messages array (simplified interface)
• Access to multiple providers through routing modes

Next Steps

Get Started

1. Create an account at TokenRouter
2. Generate an API key in the console
3. Add your provider keys (OpenAI, Anthropic, etc.)
4. Install the SDK and make your first request

Learn More

• Authentication - API key management
• Quickstart Guide - Get up and running in 5 minutes
• Provider Keys - How to add and manage provider credentials
• Routing Rules - Advanced routing configuration