> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mcp-use.com/llms.txt
> Use this file to discover all available pages before exploring further.

# LLM Integration

> Connect agents to OpenAI, Anthropic, Google, and more

The MCPAgent works with any modern LLM provider through LangChain's unified interface. Connect to OpenAI, Anthropic, Google, Groq, or any other LangChain-compatible provider that supports tool calling.

## Requirements

Your chosen LLM must support:

* **Tool calling**: Also known as function calling - required for MCP tool execution
* **Structured output**: For type-safe responses (optional but recommended)
* **Streaming**: For real-time response streaming (optional)

## Provider Integration Examples

### OpenAI

```typescript theme={null}
import { ChatOpenAI } from '@langchain/openai'
import { MCPAgent, MCPClient } from 'mcp-use'

// Initialize OpenAI model
const llm = new ChatOpenAI({
  model: "gpt-5.5",
  temperature: 0.7,
  apiKey: process.env.OPENAI_API_KEY  // Or set OPENAI_API_KEY env var
})

// Create agent
const agent = new MCPAgent({ llm, client })
```

### Anthropic

```typescript theme={null}
import { ChatAnthropic } from '@langchain/anthropic'
import { MCPAgent, MCPClient } from 'mcp-use'

// Initialize Claude model
const llm = new ChatAnthropic({
  model: 'claude-sonnet-4-6',
  temperature: 0.7,
  apiKey: process.env.ANTHROPIC_API_KEY  // Or set ANTHROPIC_API_KEY env var
})

// Create agent
const agent = new MCPAgent({ llm, client })
```

### Google Gemini

```typescript theme={null}
import { ChatGoogleGenerativeAI } from '@langchain/google-genai'
import { MCPAgent, MCPClient } from 'mcp-use'

// Initialize Gemini model
const llm = new ChatGoogleGenerativeAI({
  model: 'gemini-3.1-flash',
  temperature: 0.7,
  apiKey: process.env.GOOGLE_API_KEY  // Or set GOOGLE_API_KEY env var
})

// Create agent
const agent = new MCPAgent({ llm, client })
```

### Groq

```typescript theme={null}
import { ChatGroq } from '@langchain/groq'
import { MCPAgent, MCPClient } from 'mcp-use'

// Initialize Groq model
const llm = new ChatGroq({
  model: 'llama-4-scout-17b-16e-instruct',
  temperature: 0.7,
  apiKey: process.env.GROQ_API_KEY  // Or set GROQ_API_KEY env var
})

// Create agent
const agent = new MCPAgent({ llm, client })
```

For more LLM providers and detailed integration examples, visit the [LangChain Chat Models documentation](https://python.langchain.com/docs/integrations/chat/).