home / mcp / zendesk mcp server

Zendesk MCP Server

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "koundinya-zd-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "zd-mcp-server"
      ],
      "env": {
        "ZENDESK_EMAIL": "[email protected]",
        "ZENDESK_TOKEN": "your-zendesk-api-token",
        "ZENDESK_SUBDOMAIN": "your-company"
      }
    }
  }
}

This MCP server lets you control Zendesk Support tickets through natural language interactions with AI assistants. You can search, create, update, and manage tickets, add comments or internal notes, and handle linked incidents—all via conversational prompts. It integrates securely with Zendesk using API tokens and can be run locally or wired to various MCP clients for streamlined workflows.

How to use

You connect a client to the Zendesk MCP Server using a local or remote MCP configuration. Once connected, you can ask your AI assistant to retrieve ticket details, search for tickets with Zendesk’s query syntax, create new tickets, update existing ones, or add both public comments and private notes. You can also inspect linked incidents, manage ticket tags, and perform batch-like operations by chaining queries and actions in a single conversational flow. Start by ensuring your environment variables are set with your Zendesk credentials and subdomain, then configure your MCP client to point to the server you run (either via npx or a globally installed command). Use natural language prompts like “Show me all high priority tickets,” “Create a ticket for a login issue with urgent priority,” or “Add a public note to ticket #12345.” If you need to refine searches, leverage Zendesk’s query syntax directly within your prompts.

How to install

Prerequisites: Node.js is recommended. You will also need a Zendesk API token and your Zendesk subdomain. Ensure you have a supported MCP client ready to connect to the server.

Step 1: Install the MCP server using npm for a quick start.

Step 2: Run the MCP server locally or connect via an MCP client. Use the appropriate command shown in the setup examples below.

Step 3: Verify the server is reachable from your MCP client and test a simple query like fetching a ticket or performing a search to confirm the integration works end-to-end.

Configuration and usage notes

Set your Zendesk credentials and subdomain as environment variables in your MCP client configuration. The key variables you’ll use are ZENDESK_EMAIL, ZENDESK_TOKEN, and ZENDESK_SUBDOMAIN. These values must be provided to the MCP client so the server can authenticate requests to Zendesk on your behalf.

If you plan to use Claude Desktop or Cursor with the Zendesk MCP Server, add an MCP server entry that points to the server runner (either via npx or a globally installed command). The environment variables must be included in each client configuration to ensure proper authentication.

Common usage patterns include: creating tickets, updating statuses, adding notes, and performing complex searches. You can chain prompts to perform multiple actions in a single conversation, such as finding urgent tickets and then updating their priority or adding notes about investigations.

Troubleshooting and tips

Authentication issues often stem from an incorrect API token, a mismatched email, or an incorrect subdomain. Double-check each value and regenerate the token if necessary. Ensure API access is enabled and that the user role has sufficient permissions.

If you encounter server-not-found errors, verify that the MCP client configuration is correct and that the server runner is reachable. Use npx to run the server if you are not installing it globally. Confirm that the environment variables are set in the client’s environment before starting a session.

For debugging, enable verbose logs in your MCP client and monitor the terminal or the client’s log destination. Look for authentication or permission warnings that indicate missing scopes or expired credentials.

Advanced usage and examples

You can perform advanced searches using Zendesk’s query syntax, such as filtering by status, priority, creation date, or tags, and you can combine criteria to narrow results. You can also retrieve detailed ticket information, including comments and history, to provide comprehensive context to agents or automation.

Tools and endpoints (high-level capabilities)

The server provides tools for retrieving a ticket by ID, getting detailed ticket information with comments, performing searches, creating and updating tickets, adding private notes, adding public notes, and retrieving linked incidents. These actions enable conversational agents to manage Zendesk tickets end-to-end through natural language.

Available tools

zendesk_get_ticket

Retrieve a ticket by ID

zendesk_get_ticket_details

Get detailed ticket with comments

zendesk_search

Search tickets with query syntax

zendesk_create_ticket

Create a new ticket

zendesk_update_ticket

Update ticket properties

zendesk_add_private_note

Add internal agent notes

zendesk_add_public_note

Add public customer comments

zendesk_get_linked_incidents

Get incident tickets linked to problems