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

# MCP Server

> Search Knot's documentation from your IDE, terminal, or any MCP-compatible tool.

<Warning>
  **The Knot MCP server only searches Knot's documentation. It does not call the Knot API.** It cannot create sessions, read or write merchant account data, or take any action in your Knot account. To build an integration with the Knot API, use [skills](/ai/skills).
</Warning>

Knot provides an <Tooltip tip="An MCP server is a backend service that implements the Model Context Protocol, which is a standard that allows AI models and agents to interact with tools, data sources, and systems in a structured way.">MCP server</Tooltip> that lets you search this documentation via natural language from your IDE, terminal, or any MCP-compatible tool.

## Setup

<Steps>
  <Step title="Install the MCP server">
    Run the following command in your terminal to install the Knot MCP server as a project-level MCP locally. It auto-detects your AI tools (Cursor, Claude Code, VS Code, etc.) and configures them automatically.

    ```bash theme={"system"}
    npx add-mcp https://docs.knotapi.com/mcp --name knot-docs
    ```

    For the Claude Code Desktop App, you may also follow Claude's docs on installing an MCP server [here](https://support.claude.com/en/articles/10949351-getting-started-with-local-mcp-servers-on-claude-desktop).
  </Step>

  <Step title="Begin prompting">
    Ask questions, and your agent searches Knot's documentation to retrieve answers directly in your development environment.

    **Example prompts:**

    * "How do I integrate the Knot iOS SDK?"
    * "How can I get the available merchant list for TransactionLink?"
    * "Tell me how to create a session and what do I do after that?"
  </Step>
</Steps>