# OpenRouter MCP Server API Documentation
## Overview
This MCP server provides tools and resources for interacting with OpenRouter's unified AI model API, giving you access to 400+ AI models through a single interface.
## Resources
### Available Models (`openrouter://models`)
Returns a comprehensive list of all available models with their specifications.
**Response Format:**
```json
{
"data": [
{
"id": "openai/gpt-4",
"name": "GPT-4",
"description": "OpenAI's most capable model",
"context_length": 8192,
"pricing": {
"prompt": "0.00003",
"completion": "0.00006"
}
}
]
}
```
### Model Pricing (`openrouter://pricing`)
Returns simplified pricing information for all models.
### Usage Statistics (`openrouter://usage`)
Returns your usage statistics (placeholder - OpenRouter doesn't provide direct usage API).
## Tools
### `list_models`
Get a formatted list of available OpenRouter models.
**Parameters:** None
**Returns:** Formatted text with model information including ID, name, description, context length, and pricing.
### `chat_with_model`
Send a message to a specific OpenRouter model.
**Parameters:**
- `model` (string, required): OpenRouter model ID (e.g., "openai/gpt-4")
- `message` (string, required): Message to send to the model
- `max_tokens` (number, optional): Maximum tokens in response (default: 1000)
- `temperature` (number, optional): Temperature for response randomness (default: 0.7)
- `system_prompt` (string, optional): System prompt for the conversation
**Returns:** Model response with usage statistics.
### `compare_models`
Compare responses from multiple models using the same prompt.
**Parameters:**
- `models` (array of strings, required): Array of model IDs to compare
- `message` (string, required): Message to send to all models
- `max_tokens` (number, optional): Maximum tokens per response (default: 500)
**Returns:** Formatted comparison showing each model's response.
### `get_model_info`
Get detailed information about a specific model.
**Parameters:**
- `model` (string, required): Model ID to get information about
**Returns:** Complete model specification including capabilities, pricing, and limits.
## Error Handling
All tools and resources include comprehensive error handling:
- API key validation
- Rate limit handling
- Model availability checks
- Network error recovery
## Usage Examples
### Chat with GPT-4
```json
{
"name": "chat_with_model",
"arguments": {
"model": "openai/gpt-4",
"message": "Explain quantum computing in simple terms",
"max_tokens": 500
}
}
```
### Compare Multiple Models
```json
{
"name": "compare_models",
"arguments": {
"models": ["openai/gpt-4", "anthropic/claude-3-sonnet", "google/gemini-pro"],
"message": "What is the meaning of life?",
"max_tokens": 300
}
}
```
## Configuration
Required environment variables:
- `OPENROUTER_API_KEY`: Your OpenRouter API key
- `OPENROUTER_BASE_URL`: OpenRouter API base URL (default: https://openrouter.ai/api/v1)
- `OPENROUTER_SITE_URL`: Your site URL for OpenRouter attribution
- `OPENROUTER_APP_NAME`: Your app name for OpenRouter attribution
