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

# Using MCP Servers

LangBot supports using tools from MCP servers. You can add MCP servers to leverage the tool calling capabilities in large language models for better handling of LLM requests.

<Info title="- MCP service supports HTTP Streamable, SSE and stdio modes.">
  * MCP servers recommended acquisition channels:
    * [ModelScope](https://modelscope.cn/mcp)
    * [https://mcp.so](https://mcp.so)
    * [https://composio.dev/toolkits](https://composio.dev/toolkits)
</Info>

## Enable Function Calling for Models

As shown in the image, enable the `Function Calling` feature for the model.

<img width="400px" src="https://mintcdn.com/langbot/OB26CZsqXqbSrW_A/images/zh/deploy/mcp/mcp_01.png?fit=max&auto=format&n=OB26CZsqXqbSrW_A&q=85&s=1a9e0e6282a111e010717443e01032db" alt="enable function call" data-path="images/zh/deploy/mcp/mcp_01.png" />

## Create MCP Server

Click `Plugin Management`, then `MCP Management`, as shown:

<img width="1000px" src="https://mintcdn.com/langbot/OB26CZsqXqbSrW_A/images/zh/deploy/mcp/mcp02.png?fit=max&auto=format&n=OB26CZsqXqbSrW_A&q=85&s=ebf0a75f42cc0824e8d132695a41e155" alt="server page" data-path="images/zh/deploy/mcp/mcp02.png" />

Click `Add` in the top right corner, then `Create MCP Server`,

<img width="400px" src="https://mintcdn.com/langbot/OB26CZsqXqbSrW_A/images/zh/deploy/mcp/mcp03.png?fit=max&auto=format&n=OB26CZsqXqbSrW_A&q=85&s=1cd364486aaa1573f05bc3a5161a56aa" alt="create new mcp server" data-path="images/zh/deploy/mcp/mcp03.png" />

For example, to use the Fetch HTML server, fill in the corresponding URL, click `Test` to test the MCP server, and click `Submit` to add the server.

After adding, click the switch on the card to connect to the MCP server. Click the card to view the MCP server details:

<img width="400px" src="https://mintcdn.com/langbot/OB26CZsqXqbSrW_A/images/zh/deploy/mcp/mcp05.png?fit=max&auto=format&n=OB26CZsqXqbSrW_A&q=85&s=8f0a70cb1801e0d8ef959977416e00d0" alt="detailed server" data-path="images/zh/deploy/mcp/mcp05.png" />

## Effect

You need to select `Local Agent` as the runner in the pipeline settings and use a model that supports function calling.

<img width="1000px" src="https://mintcdn.com/langbot/OB26CZsqXqbSrW_A/images/zh/deploy/mcp/mcp04.png?fit=max&auto=format&n=OB26CZsqXqbSrW_A&q=85&s=98927bee542e9a1353ab72eb431f3d50" alt="dialog with mcp" data-path="images/zh/deploy/mcp/mcp04.png" />