Skip to main content

MCP Server

Overview

Holistics provides an MCP server that lets your own AI agents access and perform BI tasks on your Holistics workspace.

Demo

Here is a demo where Claude Desktop uses Holistics MCP to build a business report:

Why MCP server?

It enables you to combine and leverage:

  • Holistics resources and features
    • Data analytics resources: datasets, dashboards, etc.
    • Holistics features: Explore Data AI, create dashboards, export data, etc.
  • Your own AI agent tool (e.g. Claude Desktop, Cursor, Slack bot) with:
    • Its reasoning and research capabilities
    • Other "tools" that you provided to it
    • etc.

Protocol

Holistics MCP server supports:

How to connect

Prerequisites

  1. Enable Holistics AI and MCP server in your Holistics workspace settings
ai-mcp-setting
  1. Obtain a Holistics API key
  2. Obtain the MCP server address of your data center

Using remote url config

Example supported tools:

Please refer to their linked docs for more details on how to configure.

{
"mcpServers": {
"holistics": {
"url": "<MCP_SERVER_ADDRESS>",
"headers": {
"X-Holistics-Key": "<YOUR_HOLISTICS_API_KEY>"
}
}
}
}

Using local command config

Example supported tools:

Please refer to their linked docs for more details on how to configure.

{
"mcpServers": {
"holistics": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"<MCP_SERVER_ADDRESS>",
"--header",
"X-Holistics-Key:<YOUR_HOLISTICS_API_KEY>"
]
}
}
}

Notes:

  • Do not leave any spaces between X-Holistics-Key: and <YOUR_HOLISTICS_API_KEY>

Example: Claude Desktop

  1. Go to Claude Desktop Settings > Developer > Edit Config claude-desktop-mcp.png
  2. Input the above json into the config

Using OpenAI API

import OpenAI from "openai";
import type { ResponseInput } from 'openai/resources/responses/responses';

const client = new OpenAI({
apiKey: "<YOUR_OPENAI_KEY>",
});

async function chat(input: ResponseInput) {
const completion = await client.responses.create({
model: "gpt-4.1",
input: input,
tools: [
// Other tools
{
type: "web_search",
},
// BEGIN HOLISTICS MCP DECLARATION
{
type: "mcp",
server_label: "holistics_mcp",
server_url: "<MCP_SERVER_ADDRESS>",
server_description: 'Holistics MCP to access my BI and business data',
headers: {
"X-Holistics-Key": "<YOUR_HOLISTICS_API_KEY>",
},
require_approval: "never", // optional
},
// END HOLISTICS MCP DECLARATION
],
tool_choice: "auto",
parallel_tool_calls: true,
});

console.log(completion.output_text)
}

Using a Slack (MCP Client) bot

See this guide Running a self-hosted Slack bot.

Troubleshooting

Verify that Holistics MCP server is enabled

You can use MCP Inspector to test your Holistics MCP server.

ai-mcp-inspector

Steps:

  1. Run MCP Inspector (in your terminal): npx @modelcontextprotocol/inspector
    • MCP Inspector will be automatically opened in your browser
  2. Configure Holistics MCP:
    1. Choose Transport Type: Streamable HTTP
    2. Enter URL: <MCP_SERVER_ADDRESS>
    3. Configure Authentication > API Token Authentication
      • Header Name: X-Holistics-Key
      • Bearer Token: <YOUR_HOLISTICS_API_KEY>
  3. Click Connect button
    • You should see "Connected" status below the button
  4. Go to Tools tab -> Click List Tools
    • You should see the Tools listed

Open Markdown
Let us know what you think about this document :)