API Reference

The LLM Router API is fully compatible with the OpenAI API specification, allowing you to switch from OpenAI (or any other provider) with minimal or zero code changes. Base URL

https://api.llmrouter.app/v1

All requests require authentication using your LLM Router API key:

Authorization: Bearer sk_llmr_your_api_key_here

Supported Endpoints

Endpoint	Method	Description
`/chat/completions`	POST	Main chat completions endpoint (most used)
`/messages`	POST	Simplified message-based conversations
`/embeddings`	POST	Generate embeddings for text
`/images/generations`	POST	Generate images (DALL·E, Flux, etc.)
`/models`	GET	List all available models
`/models/{id}`	GET	Get details about a specific model
`/responses`	POST	Responses API (OpenAI Responses API)

Core Features

LLM Router extends the standard OpenAI format with a powerful gateway object that enables:

Intelligent model routing using tags
Automatic Skill injection (enableAutoSearch)
Zero Data Retention (ZDR)
Context compression & pruning
PII redaction
Custom routing rules

Example Request with Advanced Features

{
  "model": "anthropic/claude-opus-4.6",
  "messages": [
    { "role": "user", "content": "Write a new billing webhook handler" }
  ],
  "gateway": {
    "zdr": true,
    "skills": {
      "skillIds": ["stripe-api", "node-best-practices"],
      "enableAutoSearch": true
    },
    "tags": ["coding", "backend"]
  }
}

Philosophy

LLM Router gives you the best of both worlds:

Full OpenAI compatibility → Drop-in replacement
Powerful extensions → Smart routing, Skills, privacy, and optimization

You can start simple (just change the base URL) and gradually unlock advanced features as needed.