What is Deskmcp

DeskMCP is a Model Context Protocol server that facilitates interaction between AI agents and the Teamwork Desk API, enabling seamless management of customer support tasks.

Use cases

Use cases for DeskMCP include automating ticket creation and management, retrieving customer information, managing company and user data, and generating insightful reports.

How to use

To use DeskMCP, install it via Go or Docker, configure your Teamwork Desk API credentials, and start the server to allow AI assistants to perform actions on your Teamwork Desk account.

Key features

DeskMCP offers full support for Teamwork Desk API resources, advanced filtering capabilities, JSON response formatting for easy parsing, and Docker support for straightforward deployment.

Where to use

DeskMCP is applicable in customer support environments, particularly for businesses using Teamwork Desk to manage tickets, customers, and user interactions.

DeskMCP

An MCP (Model Context Protocol) server for the Teamwork Desk API. This server allows AI agents to interact with Teamwork Desk through a standardized protocol.

Features

• Full support for Teamwork Desk API resources:
  • Tickets
  • Customers
  • Companies
  • Users
  • Tags
  • Ticket Types
  • Ticket Statuses
• Advanced filtering capabilities for all list operations
• JSON response formatting for easy parsing
• Docker support for easy deployment

Installation

Option 1: Direct Installation

go install github.com/ready4god2513/deskmcp/cmd/mcp

Option 2: Docker Installation

1. Build the Docker image:

docker build -t deskmcp .

2. Run the container (recommended to use a secrets manager or environment file for sensitive data):

# Using environment variables (not recommended for production)
docker run -d \
  -e DESK_API_URL=https://yourcompany.teamwork.com \
  -e DESK_API_TOKEN=your_api_token \
  --name deskmcp \
  deskmcp

# Using environment file (more secure)
docker run -d \
  --env-file .env \
  --name deskmcp \
  deskmcp

3. To view logs:

docker logs -f deskmcp

4. To stop the container:

docker stop deskmcp

Configuration

The server requires the following environment variables:

• DESK_API_URL: Your Teamwork Desk API URL
• DESK_API_TOKEN: Your Teamwork Desk API token

Getting Started

What is this tool?

DeskMCP is a bridge between your Teamwork Desk account and AI assistants like Claude. It allows AI assistants to perform actions in your Teamwork Desk account, such as:

• Creating and managing support tickets
• Looking up customer information
• Managing companies and users
• Generating reports and insights

How to Use with AI Assistants

1. Install the Tool

   • Make sure you have Go installed on your computer
   • Run the installation command shown above
   • Set up your Teamwork Desk API credentials

2. Start the Server

   • Open a terminal window
   • Run the mcp command
   • Keep this terminal window open while using the tool

3. Using with Claude

   • Open Claude in your web browser
   • In your conversation, you can ask Claude to:
     • “Show me all open tickets”
     • “Create a new ticket for customer John Smith”
     • “Find all tickets from company Acme Inc”
     • “List all customers who haven’t been contacted in 30 days”

4. Tips for Best Results

   • Be specific in your requests
   • Use natural language (e.g., “show me” instead of technical commands)
   • Ask for clarification if you’re not sure what information is available
   • Request data in specific formats (e.g., “as a table” or “with dates sorted”)
   • Ask for explanations of the data when needed

5. Safety and Security

   • The tool only has access to what your API token allows
   • All actions are logged in your Teamwork Desk account
   • You can revoke access at any time by changing your API token
   • The tool runs locally on your machine, keeping your data secure

Available Tools

Tickets

• list_tickets: List all tickets with optional filters
• get_ticket: Get a specific ticket by ID
• create_ticket: Create a new ticket

Customers

• list_customers: List all customers with optional filters
• get_customer: Get a specific customer by ID
• create_customer: Create a new customer

Companies

• list_companies: List all companies with optional filters
• get_company: Get a specific company by ID
• create_company: Create a new company

Users

• list_users: List all users with optional filters
• get_user: Get a specific user by ID
• create_user: Create a new user

Tags

• list_tags: List all tags with optional filters
• get_tag: Get a specific tag by ID
• create_tag: Create a new tag

Ticket Types

• list_ticket_types: List all ticket types with optional filters
• get_ticket_type: Get a specific ticket type by ID
• create_ticket_type: Create a new ticket type

Ticket Statuses

• list_ticket_statuses: List all ticket statuses with optional filters
• get_ticket_status: Get a specific ticket status by ID
• create_ticket_status: Create a new ticket status

Filter Usage

All list operations support filtering through the filter parameter. Here are some examples:

Simple Equality Filters

{
  "filter": {
    "status": "open",
    "priority": "high"
  }
}

Multiple Conditions

{
  "filter": {
    "status": "open",
    "priority": "high",
    "customer_id": "123"
  }
}

Common Filter Fields

Tickets

• status: “open”, “pending”, “closed”, etc.
• priority: “low”, “medium”, “high”, “urgent”
• customer_id: Customer ID
• agent_id: Agent ID
• company_id: Company ID
• created_at: Date range
• updated_at: Date range

Customers

• email: Customer email
• company_id: Company ID
• created_at: Date range
• updated_at: Date range

Companies

• name: Company name
• created_at: Date range
• updated_at: Date range

Users

• email: User email
• role: User role
• created_at: Date range
• updated_at: Date range

Tags

• name: Tag name
• created_at: Date range
• updated_at: Date range

Ticket Types

• name: Ticket type name
• created_at: Date range
• updated_at: Date range

Ticket Statuses

• name: Ticket status name
• created_at: Date range
• updated_at: Date range

Use Cases

1. Customer Support Automation

• Automatically create tickets from customer emails
• Route tickets to appropriate agents based on content
• Generate responses using AI
• Track customer satisfaction metrics

2. Customer Relationship Management

• Track customer interactions across multiple channels
• Identify high-value customers
• Monitor customer satisfaction trends
• Generate customer reports

3. Team Management

• Monitor agent performance
• Track ticket resolution times
• Identify training opportunities
• Balance workload across team members

4. Business Intelligence

• Generate custom reports
• Track key performance indicators
• Analyze customer support trends
• Identify areas for improvement

5. Integration with Other Systems

• Connect with CRM systems
• Integrate with project management tools
• Sync with marketing automation platforms
• Link with accounting software

6. AI-Powered Features

• Automated ticket categorization
• Sentiment analysis of customer messages
• Smart ticket routing
• Predictive response suggestions
• Automated follow-up scheduling

Claude Desktop Configuration

To use DeskMCP with Claude Desktop:

1. Make sure DeskMCP is running locally (either through direct installation or Docker)
2. Open Claude Desktop
3. Go to Settings (gear icon)
4. Navigate to the “Tools” section
5. Add a new MCP server with the following configuration:

{
  "mcpServers": {
    "Teamwork Desk": {
      "command": "mcp",
      "args": [],
      "env": {
        "DESK_API_URL": "https://yourcompany.teamwork.com",
        "DESK_API_TOKEN": "your_api_token"
      },
      "autostart": true,
      "autorestart": true
    }
  }
}

Note: The mcp command must be in your system’s PATH. If you installed it using go install, it should be available. If you’re using Docker, you’ll need to use the appropriate Docker command instead.

6. Save the configuration
7. Restart Claude Desktop

Now you can use natural language to interact with your Teamwork Desk data through Claude. For example:

• “Show me all open tickets”
• “Create a new ticket for customer John Smith”
• “Find all tickets from company Acme Inc”

Troubleshooting

Server Disconnected Error

If you see “MCP Teamwork Desk: Server disconnected” or “spawn mcp ENOENT” error:

1. Find the location of your mcp executable:

   which mcp
   

2. Verify the full path to mcp in your configuration:

3. If which mcp returns nothing, the executable isn’t in your PATH. Add your Go bin directory to PATH:

   # Add to your ~/.zshrc or ~/.bash_profile
   export PATH=$PATH:$(go env GOPATH)/bin
   

   Then restart your terminal and Claude Desktop.

4. Alternatively, reinstall the tool:

   go install github.com/ready4god2513/deskmcp/cmd/mcp@latest
   

5. For Docker users:

   # Use the full Docker command in the configuration
   {
     "mcpServers": {
       "Teamwork Desk": {
         "command": "docker",
         "args": ["run", "--rm", "-e", "DESK_API_URL=https://yourcompany.teamwork.com", "-e", "DESK_API_TOKEN=your_api_token", "deskmcp"],
         "autostart": true,
         "autorestart": true
       }
     }
   }
   

License

MIT