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

# Quick Start

> Create your first MCP server and connect an AI client in under 5 minutes.

## Prerequisites

* WordPress 6.2+ with PHP 8.0+
* GetMCP plugin installed and activated ([Installation Guide](/docs/installation))
* An API you want to expose (or use a pre-built template for testing)

## Step 1: Create a Server

<Steps>
  <Step title="Navigate to GetMCP Dashboard">
    In your WordPress admin sidebar, click **GetMCP** to open the dashboard.

    <img src="https://mintcdn.com/infiwebs/X3D8lQwdysxDpY3z/images/after-activated-dashboard.jpeg?fit=max&auto=format&n=X3D8lQwdysxDpY3z&q=85&s=fc13ac413e487efc9d7587660b2f0a39" alt="GetMCP Dashboard after installation" width="1663" height="908" data-path="images/after-activated-dashboard.jpeg" />
  </Step>

  <Step title="Click Create Server">
    Click the **Create Server** button in the top right corner.

    <img src="https://mintcdn.com/infiwebs/X3D8lQwdysxDpY3z/images/create-server-dashboard.jpeg?fit=max&auto=format&n=X3D8lQwdysxDpY3z&q=85&s=d99c90d452f946cd019b4895f3097e72" alt="Create Server button on dashboard" width="1334" height="689" data-path="images/create-server-dashboard.jpeg" />
  </Step>

  <Step title="Fill in Server Details">
    Enter a **Name** (e.g., `My Weather`) and Description (optional). The slug and server ID are auto-generated.

    Click **Create** to save.

    <img src="https://mintcdn.com/infiwebs/X3D8lQwdysxDpY3z/images/create-server.jpeg?fit=max&auto=format&n=X3D8lQwdysxDpY3z&q=85&s=35f9dc689f5e5ee929ddb56717bb6a94" alt="Create Server form" width="1440" height="856" data-path="images/create-server.jpeg" />
  </Step>
</Steps>

## Step 2: Add a Tool

<Steps>
  <Step title="Open the Tools Tab">
    Inside your new server, click the **Tools** tab.
  </Step>

  <Step title="Click Add Tool">
    Click **Add Tool** to open the tool editor.

    <img src="https://mintcdn.com/infiwebs/gDJxCN8mnpgMsXPG/images/import-options.png?fit=max&auto=format&n=gDJxCN8mnpgMsXPG&q=85&s=5dbac5e035ea5e84a1b33ce99eede3ec" alt="Tool Creation Options" width="1406" height="658" data-path="images/import-options.png" />
  </Step>

  <Step title="Configure the Tool">
    Fill in the tool details:

    * **Name**: `get_current_weather` (use snake\_case — this is how AI clients call it)
    * **Description**: `Get real-time weather conditions for a location by coordinates`
    * **Endpoint URL**: `https://api.open-meteo.com/v1/forecast`
    * **HTTP Method**: `GET`

    <img src="https://mintcdn.com/infiwebs/X3D8lQwdysxDpY3z/images/tool-editor.png?fit=max&auto=format&n=X3D8lQwdysxDpY3z&q=85&s=701ccdcbec9691148ac932b4a09c9834" alt="Tool Configuration" width="1683" height="903" data-path="images/tool-editor.png" />
  </Step>

  <Step title="Define Input Parameters">
    In the **Input Schema** section, add three parameters:

    | Name              | Type      | Required | Description                          | Mapping |
    | ----------------- | --------- | -------- | ------------------------------------ | ------- |
    | `latitude`        | `number`  | Yes      | `Latitude of the location`           | Query   |
    | `longitude`       | `number`  | Yes      | `Longitude of the location`          | Query   |
    | `current_weather` | `boolean` | No       | `Include current weather conditions` | Query   |

    All three map as **Query** parameters — they are appended to the URL as `?latitude=...&longitude=...&current_weather=true`.
  </Step>

  <Step title="Save the Tool">
    Click **Save Tool**. Your tool is now configured.
  </Step>
</Steps>

## Step 3: Test the Tool

Before connecting an AI client, verify the tool works correctly.

<Steps>
  <Step title="Open the Test Panel">
    On the tool row, click the **Test** button.
  </Step>

  <Step title="Enter Test Arguments">
    In the test panel, enter the following as the arguments (London coordinates):

    ```json theme={null}
    {
      "latitude": 51.5085,
      "longitude": -0.1257,
      "current_weather": true
    }
    ```
  </Step>

  <Step title="Run the Test">
    Click **Run Test**. You should see the weather data returned from the API.

    <img src="https://mintcdn.com/infiwebs/X3D8lQwdysxDpY3z/images/tool-test-modal.png?fit=max&auto=format&n=X3D8lQwdysxDpY3z&q=85&s=91ab7ce67639fd2f68c954ff31a4e61e" alt="Tool Test Panel" width="1052" height="906" data-path="images/tool-test-modal.png" />
  </Step>
</Steps>

## Step 4: Connect an AI Client

<Steps>
  <Step title="Copy the Server URL">
    On the server detail page, find the **Server URL** shown at the top. It looks like:

    ```
    https://yoursite.com/mcp/{server-slug}
    ```

    Copy this URL. You can also click **How to connect** on the server detail page to get a pre-filled config snippet for your client.
  </Step>

  <Step title="Configure Your AI Client">
    <img src="https://mintcdn.com/infiwebs/X3D8lQwdysxDpY3z/images/claude-config.png?fit=max&auto=format&n=X3D8lQwdysxDpY3z&q=85&s=8840539ad4dbd5ff56c5f84d4066aec2" alt="AI Client Configuration" width="1112" height="450" data-path="images/claude-config.png" />

    <Tabs>
      <Tab title="Claude Desktop">
        Open Developer Settings and click **Edit Config**:

        * **macOS:** Menu bar → Claude → Settings → Developer → Edit Config
        * **Windows:** Breadcrumb menu (top bar) → File → Settings → Developer → Edit Config

        **macOS / Linux:**

        ```json theme={null}
        {
          "mcpServers": {
            "{server-slug}": {
              "command": "npx",
              "args": [
                "mcp-remote",
                "https://yoursite.com/mcp/{server-slug}"
              ]
            }
          }
        }
        ```

        **Windows:**

        ```json theme={null}
        {
          "mcpServers": {
            "{server-slug}": {
              "command": "cmd",
              "args": [
                "/c",
                "npx",
                "mcp-remote",
                "https://yoursite.com/mcp/{server-slug}"
              ]
            }
          }
        }
        ```

        Fully quit and restart Claude Desktop — closing the window is not enough.
      </Tab>

      <Tab title="Claude Code">
        Open your terminal and run:

        ```bash theme={null}
        claude mcp add --transport http {server-slug} https://yoursite.com/mcp/{server-slug}
        ```

        The server is added immediately — no restart needed. Run `/mcp` to verify it appears.
      </Tab>

      <Tab title="Cursor">
        Go to **Settings → Tools & MCP → Add a custom MCP**. In the `mcp.json` file that opens, add:

        ```json theme={null}
        {
          "mcpServers": {
            "{server-slug}": {
              "url": "https://yoursite.com/mcp/{server-slug}"
            }
          }
        }
        ```

        Restart Cursor after saving.
      </Tab>

      <Tab title="Windsurf">
        Go to **Settings → Cascade → Open MCP Registry**. In the `mcp_config.json` file that opens, add:

        ```json theme={null}
        {
          "mcpServers": {
            "{server-slug}": {
              "serverUrl": "https://yoursite.com/mcp/{server-slug}"
            }
          }
        }
        ```

        Restart Windsurf after saving.
      </Tab>

      <Tab title="VS Code">
        Open (or create) `.vscode/mcp.json` in your project root and add:

        ```json theme={null}
        {
          "servers": {
            "{server-slug}": {
              "type": "http",
              "url": "https://yoursite.com/mcp/{server-slug}"
            }
          }
        }
        ```

        Save — VS Code detects the file automatically, no restart needed.
      </Tab>

      <Tab title="OpenAI Playground">
        1. Go to [platform.openai.com/playground/prompts](https://platform.openai.com/playground/prompts)
        2. Click the **tools icon** in the left panel → **Add MCP server**
        3. Enter your server URL and a label name, then click **Connect**

        <img src="https://mintcdn.com/infiwebs/uvEl3_c3deYVxCrc/images/openai-playground.png?fit=max&auto=format&n=uvEl3_c3deYVxCrc&q=85&s=cb5459116534c41fcee80125b70b435a" alt="OpenAI Playground MCP Connection" width="494" height="782" data-path="images/openai-playground.png" />
      </Tab>
    </Tabs>
  </Step>
</Steps>

## Step 5: Use It

Open your AI client and ask it to use your tool:

> "What's the current weather and 3-days forecast for Pune, Maharashtra, India?"

> The AI will automatically discover and call your `get_current_weather` tool, passing the coordinates as query parameters.

<img src="https://mintcdn.com/infiwebs/X3D8lQwdysxDpY3z/images/claude-desktop-response-1.png?fit=max&auto=format&n=X3D8lQwdysxDpY3z&q=85&s=053035001e5a18695180f0f3f9f68fca" alt="AI Client Calling Tool" width="1000" height="791" data-path="images/claude-desktop-response-1.png" />

<img src="https://mintcdn.com/infiwebs/X3D8lQwdysxDpY3z/images/claude-desktop-response-2.png?fit=max&auto=format&n=X3D8lQwdysxDpY3z&q=85&s=215b664f5d15e282f17947267499d714" alt="AI Client Response" width="1029" height="916" data-path="images/claude-desktop-response-2.png" />

## Next Steps

<CardGroup cols={2}>
  <Card title="Authentication" icon="lock" href="/docs/guides/authentication">
    Secure your tools with API keys and Bearer tokens
  </Card>

  <Card title="Resources & Prompts" icon="file-text" href="/docs/concepts/resources">
    Expose static content and reusable prompt templates
  </Card>

  <Card title="Analytics" icon="chart-bar" href="/docs/guides/analytics">
    Monitor tool usage and performance
  </Card>

  <Card title="cURL Import" icon="terminal" href="/docs/guides/curl-import">
    Import existing API calls directly from cURL commands
  </Card>
</CardGroup>