<!-- description: An MCP client is an AI agent or application that connects to MCP servers to access tools, resources, and prompts using the Model Context Protocol. -->

# MCP Client

An **MCP client** is an AI agent or application that connects to [MCP servers](mcp-server) to access tools, resources, and prompts through the [Model Context Protocol](mcp). The client is the consumer side of the MCP architecture — it initiates connections and calls capabilities that servers expose.

## Role of the MCP Client

The MCP client is responsible for:

- Discovering available tools, resources, and prompts from connected servers
- Calling tools on behalf of the AI model during task execution
- Managing the lifecycle of server connections
- Routing tool calls to the appropriate server

The most common MCP client today is [Claude Code](claude-code), Anthropic's agentic coding assistant.

## How MCP Clients Discover Tools

When an MCP client connects to a server, it calls `tools/list` to enumerate all available tools. The server responds with tool names, descriptions, and JSON schemas for their parameters. The client passes this information to the LLM so it knows what tools are available during inference.

## MCP Client Configuration

MCP clients are configured with a list of servers to connect to. In Claude Code, this is done via the `.claude/mcp.json` config file:

```json
{
  "mcpServers": {
    "agentrq": {
      "type": "http",
      "url": "https://mcp.agentrq.com/YOUR_WORKSPACE_ID"
    }
  }
}
```

## Examples of MCP Clients

- **Claude Code** — Anthropic's terminal-based agentic coding assistant
- **Claude Desktop** — Anthropic's desktop chat application with MCP support
- **Custom agents** — Any application built on the Anthropic API that implements MCP

## Related Terms

- [MCP](mcp)
- [MCP Server](mcp-server)
- [MCP Tool](mcp-tool)
- [Claude Code](claude-code)
- [Tool Use](tool-use)