# Quick Start

## Quick Start Guide: Using SigilAI Cloud MCP Server

This guide helps you quickly connect your AI development tools and assistants to the SigilAI cloud-hosted Model Context Protocol (MCP) server. SigilAI provides a managed MCP service, allowing you to easily integrate powerful security scanning into your workflows.

<figure><img src="/files/lw5v5PCL3Cwymym09UR8" alt="" width="563"><figcaption></figcaption></figure>

### Prerequisites

Before you begin, ensure you have:

1. **API Key**: Obtain your API key from the SigilAI Telegram bot at <https://t.me/SigilAIBot>. This key is essential for authenticating with the MCP server.

<figure><img src="/files/jKSt8LlyIRX9ot4KLTPH" alt="" width="375"><figcaption></figcaption></figure>

2. **MCP Server Endpoint**: The SigilAI MCP Server URL is `https://mcp.sigilai.io/sse` .
3. **Compatible AI Client**: An AI assistant or development tool that supports MCP integration, such as Cline, Cursor, or Claude Desktop.

### Connecting Your AI Client to SigilAI MCP Server

<figure><img src="/files/Mux8lkoBc5Lk5zHDCfwH" alt="" width="375"><figcaption></figcaption></figure>

The SigilAI MCP server allows various AI clients to access its security tools. Below are configuration examples for popular clients. The general principle involves adding SigilAI as an MCP server provider within your client's settings, using your API key and the server endpoint.

{% tabs %}
{% tab title="Cline" %}
To configure SigilAI MCP server access within Cline:

1. **Access MCP Settings**:
   * You (as the user interacting with Cline) would typically inform Cline: "Please add a new MCP server configuration."
   * Cline would then guide you or use its internal capabilities to update its MCP server list.
2. **Provide SigilAI MCP Server Details**:
   \
   When prompted by Cline, or by editing Cline's MCP configuration file (if accessible to you as a user/developer of Cline), provide the following:
   * **Server Name**: `SigilAI` (or any preferred display name)
   * **Server URL / Endpoint**: `https://mcp.sigilai.io/sse` (or your specific endpoint)
   * **Authentication Method**: Bearer Token
   * **API Key / Token**: Your SigilAI API Key
   * **(Optional) Instructions/Description**: You can copy the server instructions from the SigilAI MCP server if your client supports displaying them.
3. **Save and Activate**:
   * Confirm the settings. Cline should then be able to connect to the SigilAI MCP server and list its available tools (e.g., `scan_url`, `scan_file`).
4. **Usage Example with Cline**:

   ```
   User: Cline, please scan the URL https://example-suspicious.com using SigilAI.

   Cline: (Internally makes an MCP tool_call to SigilAI's scan_url tool)
          Okay, I've scanned https://example-suspicious.com using SigilAI. 
          The result indicates it is potentially malicious due to [reason]. 
          It's advisable not to visit this site.
   ```

{% endtab %}

{% tab title="Cursor IDE" %}
Cursor is an AI-first code editor. To configure an external MCP server like SigilAI (assuming Cursor supports generic MCP server integration):

1. **Open Cursor Settings**:
   * Navigate to Cursor's settings/preferences panel. This is often accessible via `File > Settings` or a shortcut like `Ctrl+,` / `Cmd+,`.
2. **Locate AI / MCP / Tool Settings**:
   * Look for a section related to "AI Features," "External Tools," "Model Context Protocol," or "Custom LLM Tools." The exact naming may vary.
3. **Add New MCP Server**:
   * There might be an option to "Add MCP Server," "Configure Tool Provider," or similar.
   * Enter the following details:
     * **Name**: `SigilAI Security Scanner`
     * **Endpoint URL**: `https://mcp.sigilai.io/sse` (or your specific endpoint)
     * **Authentication Type**: Select "Bearer Token" or "API Key."
     * **API Key**: Paste your SigilAI API Key.
     * **(Optional) Description**: "Security scanning tools from SigilAI."
4. **Save Configuration**:
   * Apply or save the settings. Cursor should now attempt to connect to the SigilAI MCP server.
5. **Usage Example within Cursor**:
   * While editing code, you might invoke an AI chat or command:

     ```
     User: @SigilAI scan this file for vulnerabilities.
     ```
   * Cursor, using its AI capabilities and the configured MCP connection, would send the file content to SigilAI's `scan_file` tool and display the results.

*(Note: Specific steps for Cursor depend on its current MCP integration features. Refer to Cursor's official documentation for the most accurate instructions.)*
{% endtab %}

{% tab title="Claude Desktop App" %}
If the Claude Desktop application supports generic MCP server integration:

1. **Access Settings**:
   * Open the Claude Desktop application.
   * Look for a settings, preferences, or extensions menu. This could be under a "File" menu, a gear icon, or your profile.
2. **Find MCP / External Tools Configuration**:
   * Search for options like "Connected Services," "Tool Integrations," or "Model Context Protocol Servers."
3. **Add SigilAI MCP Server**:
   * Click an "Add New Server" or similar button.
   * Input the SigilAI server details:
     * **Server Identifier**: `SigilAI_Cloud`
     * **MCP Endpoint**: `https://mcp.sigilai.io/sse` (or your specific endpoint)
     * **Authentication**: Choose "Bearer Token" or "API Key" and provide your SigilAI API Key.
     * **Display Name**: `SigilAI Security Tools`
4. **Apply and Verify**:
   * Save the configuration. The Claude Desktop app might attempt to connect and list available tools from SigilAI.
5. **Usage Example with Claude Desktop**:
   * In a chat with Claude:

     ```
     User: Claude, can you check if this URL is safe using SigilAI: http://phishy-link.info

     Claude: (Leveraging the SigilAI MCP integration) 
             I've checked http://phishy-link.info with SigilAI. 
             It's flagged as a high-risk phishing site. I advise against visiting it.
     ```

*(Note: Specific steps for Claude Desktop depend on its MCP integration capabilities. Always consult the official Claude Desktop documentation.)*
{% endtab %}
{% endtabs %}

### Verifying the Connection

Once configured, your AI client should ideally indicate a successful connection or list the tools available from the SigilAI MCP server (e.g., `scan_url`, `scan_file`).

You can perform a simple test by asking your AI client to use one of SigilAI's tools:

* **Ask to scan a known safe URL**:
  `"Please scan the URL https://www.google.com using SigilAI."`
  \
  The expected response should indicate the URL is safe.
* **Ask to scan a (non-malicious) test file content (if applicable to your client)**:
  `"Please scan this JavaScript code for issues using SigilAI: function greet() { console.log('Hello'); }"`
  \
  The expected response should indicate no critical vulnerabilities.

### Next Steps

With your AI client successfully connected to the SigilAI Cloud MCP Server, you can now:

* **Integrate into Workflows**: Start using SigilAI's security scanning tools within your AI-assisted development, research, or content analysis tasks.
* **Review API Reference**: For more technical details on tool invocation, see the API Reference.

If you encounter issues, consult the[ FAQ & Troubleshooting](/sigilai-docs/resources/faq.md) page or contact SigilAI support.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://sigilai.gitbook.io/sigilai-docs/introduction/quick-start.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.