DigitalGenius MCP Server

Overview

The MCP Server connects your AI assistant directly to DigitalGenius via the Model Context Protocol (MCP) - an open standard for connecting AI assistants to external data and tools.

Instead of jumping between dashboards and browser tabs, you can now inspect executions, debug flows and query analytics using plain English from the likes of Claude and Gemini.

What You Can Do

Example Prompts

• Ticket 12345 got a wrong response. What happened?
• Show me all failed executions for flow X in the last 24 hours.
• What is wrong with this flow?
• Do a breakdown of CSAT per intent. Follow-up: Sample some tickets with bad CSAT for "Where is my order". what are the issues the customers are facing?
• Give me the conversation transcript and flow trail for ticket 90412. The customer says the bot gave a wrong answer.
• Pull the full nested flow tree for execution abc-123 - I need to see where it branched.
• What were the top 10 predicted intents last month?
• Compare ticket volume week-over-week for the last 8 weeks, broken down by brand.

How It Works

1. You send a natural-language message to your AI assistant (e.g. Claude).
2. The AI identifies the appropriate MCP tool(s) to call based on your intent.
3. The MCP server authenticates the request and queries DigitalGenius APIs on your behalf.
4. Results are returned to the AI, which synthesises a clear, human-readable answer for you.

Getting Started

Installing the MCP is straightforward and should only take a few minutes.

Prerequisites

1. Login to your DigitalGenius Dashboard and browse to the Settings area by clicking on the cog icon in the lower left hand corner of the page.
2. Select API Keys in the left hand menu.
3. Give your API key a name and select All in the Access Type dropdown. Click on the Create API Key button.
4. Click on the 3 dots menu for the API Key you just created and click on the Copy Encoded Token option. This will copy an authentication token to your clipboard. Store this somewhere safe.

Installing the MCP into your AI Assistant

See below for details on how to install the DigitalGenius MCP Server into your chosen AI Assistant.

Claude Desktop

1. Ensure that you have Claude Desktop and Node.js LTS installed.

2. Open the Claude Desktop app and go to Settings -> Developer -> Edit Config

3. Open claude_desktop_config.json with a text editor and add the following to the existing mcpServers section (or create one if it does not exist).

   Replace {{AUTH_TOKEN}} with the authorization token you copied from the dashboard in step 2 of the Prerequisites section above. Also replace {{ENV}} with either eu or us, depending on where your DigitalGenius tenant is hosted.

{
  "mcpServers": {
   "digitalgenius": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://flow-server.{{ENV}}.dgdeepai.com/mcp",
        "--header",
        "Authorization: Basic {{AUTH_TOKEN}}"
      ]
    }
  }
}

4. Restart Claude Desktop.
5. Verify the connection by asking List the first ten flows from my DigitalGenius org.

Claude Code

1. Ensure that you have Claude Code installed.
2. Run the following command in your terminal and ensure you replace {{AUTH_TOKEN}} with the authorization token you copied from the dashboard in step 2 of the Prerequisites section above. Also replace {{ENV}} with either eu or us, depending on where your DigitalGenius tenant is hosted.
3. 
   

   claude mcp add --scope project \
     --transport http digitalgenius "https://flow-server.{{ENV}}.dgdeepai.com/mcp", \
     --header "Authorization: Basic {{AUTH_TOKEN}}"

Gemini CLI

1. Ensure that you have Gemini CLI installed and configured and Node.js LTS installed.

2. Open your Gemini CLI settings file at ~/.gemini/settings.json.

3. Add the following to the existing mcpServers section (or create one if it does not exist).

   Ensure you replace {{AUTH_TOKEN}} with the authorization token you copied from the dashboard in step 2 of the Prerequisites section above. Also replace {{ENV}} with either eu or us, depending on where your DigitalGenius tenant is hosted.

   {
     "mcpServers": {
      "digitalgenius": {
         "command": "npx",
         "args": [
           "mcp-remote",
           "https://flow-server.{{ENV}}.dgdeepai.com/mcp",
           "--header",
           "Authorization: Basic {{AUTH_TOKEN}}"
         ]
       }
     }
   }

4. Restart Gemini CLI.

5. Verify the connection by asking List the first ten flows from my DigitalGenius org.

Other Compatible MCP clients

The DigtialGenius MCP server is compatible with any AI assistant that supports the Model Context Protocol. The guides above cover some common clients, but the same connection will work with other MCP-compatible tools if you use the following MCP server URL and auth:

Url
Ensure that you replace {{ENV}} with either eu or us depending on where your DigitalGenius tenant is hosted.

https://flow-server.{{ENV}}.dgdeepai.com/mcp

Authentication

Ensure that you pass the auth token using basic auth. Replace{{AUTH_TOKEN}} part with the authorization token you copied from the dashboard in step 2 of the Prerequisites section above.

"Authorization: Basic {{AUTH_TOKEN}}"