How to use DeskDay MCP

MCP (Model Context Protocol) Usage Guide

Overview

MCP (Model Context Protocol) is a standardized protocol that enables AI assistants to access external tools, data sources, and services. This guide covers how to set up and use MCP across different platforms and clients.

Claude Desktop

Installation

  1. Download Claude Desktop for your operating system
  2. Install and launch the application

Configuration

  1. Open the settings menu (gear icon)
  2. Navigate to Developer section
  3. Click Edit Config to open the MCP configuration file
  4. Add MCP servers in JSON format:

{
  "mcpServers": {
    "deskday": {
      "type": "http",
      "url": "https://developer.deskday.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_DESKDAY_API_KEY"
      }
    }
  }
}
  1. Save and restart Claude Desktop

Usage

  • MCP tools appear automatically in Claude's interface
  • Use them naturally in conversations without special syntax
  • Tools are context-aware and suggested when relevant

Claude Web

Custom MCP connectors in claude web require a Claude Pro or Team plan.

Step 1 — Go to Settings

  1. Click your profile icon (top right)
  2. Select Settings → Connectors
    Step 2 — Add a custom connector
  3. Click Add custom connector (or Add MCP server)
  4. Enter the following:
  5. Name: DeskDay
  6. URL: [https://[developer.deskday.ai/mcp](https://developer.deskday.ai/mcp)](https://deskday.readme.io/mcp)
  7. Click Save

    You might be prompted for an API key during conversations. Paste the Public API key generated inside Deskday control center (dd_pkxxxxx) when prompted.

    Security note: Claude.ai and ChatGPT require you to paste your API key directly into the chat to configure the MCP connection. This may expose your key in conversation history. We recommend using other platforms that support secure credential configuration without requiring your key to be entered in chat. If you are using Claude.ai or ChatGPT, please rotate your API key regularly and avoid reusing keys across integrations.

Cursor

Setup

  1. Install Cursor code editor
  2. Open SettingsFeaturesMCP
  3. Enable MCP support

Configuration

  1. Create or edit .cursor/mcp.json in your project root:
{
  "mcpServers": {
    "deskday": {
      "type": "http",
      "url": "https://developer.deskday.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_DESKDAY_API_KEY"
      }
    }
  }
}
  1. Restart Cursor

Usage

  • Access MCP tools through the command palette (Cmd/Ctrl + Shift + P)
  • Type MCP: to see available commands
  • Tools integrate with Cursor's AI features for code assistance

Visual Studio Code

Installation

  1. Install VS Code from code.visualstudio.com
  2. Install the MCP Client extension from the Extensions marketplace

Configuration

  1. Open VS Code settings (Cmd/Ctrl + ,)
  2. Search for "MCP"
  3. Configure MCP servers in settings.json:
{
  "mcpServers": {
    "deskday": {
      "type": "http",
      "url": "https://developer.deskday.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_DESKDAY_API_KEY"
      }
    }
  }
}
  1. Reload VS Code

Usage

  • Access MCP tools via the MCP sidebar panel
  • Use the command palette for quick access
  • Integrate with VS Code's built-in AI features

ChatGPT

Setup

  1. Log in to ChatGPT
  2. Ensure you have ChatGPT Plus or Enterprise subscription
  3. Access SettingsBeta Features
  4. Enable MCP Support (if available in your region)

Configuration

  1. Go to SettingsIntegrations
  2. Add MCP server endpoints:
    • Provide server URL
    • Add authentication credentials if required
  3. Save configuration

Usage

  • MCP tools appear as available actions in conversations
  • Reference tools by name in your prompts
  • ChatGPT automatically routes requests to appropriate MCP servers

Security note: Claude.ai and ChatGPT require you to paste your API key directly into the chat to configure the MCP connection. This may expose your key in conversation history. We recommend using other platforms that support secure credential configuration without requiring your key to be entered in chat. If you are using Claude.ai or ChatGPT, please rotate your API key regularly and avoid reusing keys across integrations.

GitHub Copilot / Codex

Setup

  1. Install GitHub Copilot extension in your IDE
  2. Authenticate with your GitHub account

Configuration

  1. In your IDE settings, locate Copilot configuration
  2. Add MCP server details:
    • Server endpoint URL
    • API keys or authentication tokens
  3. Save and restart your IDE

Usage

  • Copilot suggests MCP tool usage in code comments
  • Use inline chat to request tool execution
  • Tools enhance code generation and debugging capabilities

Generic MCP Client Setup

Requirements

  • Node.js 16+ or Python 3.8+
  • MCP server running and accessible
  • Network connectivity (for remote servers)

Basic Configuration Template

{
  "mcpServers": {
    "deskday": {
      "type": "http",
      "url": "https://developer.deskday.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_DESKDAY_API_KEY"
      }
    }
  }
}

Common Environment Variables

  • MCP_SERVER_URL - Server endpoint
  • MCP_API_KEY - Authentication token
  • MCP_TIMEOUT - Request timeout in milliseconds
  • MCP_DEBUG - Enable debug logging

Best Practices

  • Security: Never commit API keys to version control; use environment variables
  • Testing: Test MCP connections before production use
  • Monitoring: Enable logging to troubleshoot connection issues
  • Updates: Keep MCP servers and clients updated for latest features
  • Documentation: Reference your MCP server's documentation for available tools

Troubleshooting

Connection Issues

  • Verify server is running and accessible
  • Check firewall and network settings
  • Review authentication credentials
  • Enable debug logging for detailed error messages

Tool Not Appearing

  • Restart the client application
  • Verify MCP configuration syntax
  • Check server logs for errors
  • Ensure tool is properly exported from server

Performance Problems

  • Monitor server resource usage
  • Increase timeout values if needed
  • Reduce number of concurrent requests
  • Check network latency


Did this page help you?