Qlik Sense MCP Server

Model Context Protocol (MCP) server for integration with Qlik Sense Enterprise APIs. Provides unified interface for Repository API and Engine API operations through MCP protocol.

Table of Contents

• Overview
• Features
• Installation
• Configuration
• Usage
• API Reference
• Architecture
• Development
• Troubleshooting
• License

Overview

Qlik Sense MCP Server bridges Qlik Sense Enterprise with systems supporting Model Context Protocol. Server provides 10 comprehensive tools for complete Qlik Sense analytics workflow including application discovery, data analysis, script extraction, and metadata management.

Key Features

• Unified API: Single interface for Qlik Sense Repository and Engine APIs
• Security: Certificate-based authentication support
• Performance: Optimized queries and direct API access
• Analytics: Advanced data analysis and hypercube creation
• Metadata: Comprehensive application and field information

Features

Available Tools

Installation

Quick Start with uvx (Recommended)

The easiest way to use Qlik Sense MCP Server is with uvx:

uvx qlik-sense-mcp-server

This command will automatically install and run the latest version without affecting your system Python environment.

Alternative Installation Methods

From PyPI

pip install qlik-sense-mcp-server

From Source (Development)

git clone https://github.com/bintocher/qlik-sense-mcp.git
cd qlik-sense-mcp
make dev

System Requirements

• Python 3.12+
• Qlik Sense Enterprise
• Valid certificates for authentication
• Network access to Qlik Sense server (ports 4242 Repository, 4747 Engine)
• Ensure your MCP client model can handle large JSON responses; prefer small limits in requests during testing

Setup

1. Setup certificates

mkdir certs
# Copy your Qlik Sense certificates to certs/ directory

1. Create configuration

cp .env.example .env
# Edit .env with your settings

Configuration

Environment Variables (.env)

# Server connection
QLIK_SERVER_URL=https://your-qlik-server.company.com
QLIK_USER_DIRECTORY=COMPANY
QLIK_USER_ID=your-username

# Certificate paths (absolute paths)
QLIK_CLIENT_CERT_PATH=/path/to/certs/client.pem
QLIK_CLIENT_KEY_PATH=/path/to/certs/client_key.pem
QLIK_CA_CERT_PATH=/path/to/certs/root.pem

# API ports (standard Qlik Sense ports)
QLIK_REPOSITORY_PORT=4242
QLIK_ENGINE_PORT=4747

# Optional HTTP port for metadata requests
QLIK_HTTP_PORT=443

# SSL settings
QLIK_VERIFY_SSL=false

Optional Environment Variables

# Logging level (default: INFO)
LOG_LEVEL=INFO

# Engine WebSocket timeouts and retries
QLIK_WS_TIMEOUT=8.0     # seconds
QLIK_WS_RETRIES=2       # number of endpoints to try

MCP Configuration

Create mcp.json file for MCP client integration:

{
  "mcpServers": {
    "qlik-sense": {
      "command": "uvx",
      "args": ["qlik-sense-mcp-server"],
      "env": {
        "QLIK_SERVER_URL": "https://your-qlik-server.company.com",
        "QLIK_USER_DIRECTORY": "COMPANY",
        "QLIK_USER_ID": "your-username",
        "QLIK_CLIENT_CERT_PATH": "/absolute/path/to/certs/client.pem",
        "QLIK_CLIENT_KEY_PATH": "/absolute/path/to/certs/client_key.pem",
        "QLIK_CA_CERT_PATH": "/absolute/path/to/certs/root.pem",
        "QLIK_REPOSITORY_PORT": "4242",
        "QLIK_PROXY_PORT": "4243",
        "QLIK_ENGINE_PORT": "4747",
        "QLIK_HTTP_PORT": "443",
        "QLIK_VERIFY_SSL": "false",
        "QLIK_HTTP_TIMEOUT": "10.0",
        "QLIK_WS_TIMEOUT": "8.0",
        "QLIK_WS_RETRIES": "2",
        "LOG_LEVEL": "INFO"
      },
      "disabled": false,
      "autoApprove": [
        "get_apps",
        "get_app_details",
        "get_app_script",
        "get_app_field_statistics",
        "engine_create_hypercube",
        "get_app_field",
        "get_app_variables",
        "get_app_sheets",
        "get_app_sheet_objects",
        "get_app_object"
      ]
    }
  }
}

Environment Variables Configuration

The server requires the following environment variables for configuration:

Required Variables

• QLIK_SERVER_URL - Qlik Sense server URL (e.g., https://qlik.company.com)
• QLIK_USER_DIRECTORY - User directory for authentication (e.g., COMPANY)
• QLIK_USER_ID - User ID for authentication (e.g., your-username)

Certificate Configuration (Required for production)

• QLIK_CLIENT_CERT_PATH - Absolute path to client certificate file (.pem format)
• QLIK_CLIENT_KEY_PATH - Absolute path to client private key file (.pem format)
• QLIK_CA_CERT_PATH - Absolute path to CA certificate file (.pem format). If not specified, SSL certificate verification will be disabled

Network Configuration

• QLIK_REPOSITORY_PORT - Repository API port (default: 4242)
• QLIK_PROXY_PORT - Proxy API port for authentication (default: 4243)
• QLIK_ENGINE_PORT - Engine API port for WebSocket connections (default: 4747)
• QLIK_HTTP_PORT - HTTP API port for metadata requests (optional, only used for /api/v1/apps/{id}/data/metadata endpoint)

SSL and Security

• QLIK_VERIFY_SSL - Verify SSL certificates (true/false, default: true)

Timeouts and Performance

• QLIK_HTTP_TIMEOUT - HTTP request timeout in seconds (default: 10.0)
• QLIK_WS_TIMEOUT - WebSocket connection timeout in seconds (default: 8.0)
• QLIK_WS_RETRIES - Number of WebSocket connection retry attempts (default: 2)

Logging

• LOG_LEVEL - Logging level (DEBUG, INFO, WARNING, ERROR, default: INFO)

Usage

Start Server

# Using uvx (recommended)
uvx qlik-sense-mcp-server

# Using installed package
qlik-sense-mcp-server

# From source (development)
python -m qlik_sense_mcp_server.server

Example Operations

Get Applications List

# Via MCP client - get first 50 apps (default)
result = mcp_client.call_tool("get_apps")
print(f"Showing {result['pagination']['returned']} of {result['pagination']['total_found']} apps")

# Search for specific apps
result = mcp_client.call_tool("get_apps", {
    "name_filter": "Sales",
    "limit": 10
})

# Get more apps (pagination)
result = mcp_client.call_tool("get_apps", {
    "offset": 50,
    "limit": 50
})

Analyze Application

# Get comprehensive app analysis
result = mcp_client.call_tool("get_app_details", {
    "app_id": "your-app-id"
})
print(f"App has {len(result['data_model']['tables'])} tables")

Create Data Analysis Hypercube

# Create hypercube for sales analysis
result = mcp_client.call_tool("engine_create_hypercube", {
    "app_id": "your-app-id",
    "dimensions": ["Region", "Product"],
    "measures": ["Sum(Sales)", "Count(Orders)"],
    "max_rows": 1000
})

Get Field Statistics

# Get detailed field statistics
result = mcp_client.call_tool("get_app_field_statistics", {
    "app_id": "your-app-id",
    "field_name": "Sales"
})
print(f"Average: {result['avg_value']['numeric']}")

API Reference

get_apps

Retrieves comprehensive list of Qlik Sense applications with metadata, pagination and filtering support.

Parameters:

• limit (optional): Maximum number of apps to return (default: 50, max: 1000)
• offset (optional): Number of apps to skip for pagination (default: 0)
• name_filter (optional): Filter apps by name (case-insensitive partial match)
• app_id_filter (optional): Filter by specific app ID/GUID
• include_unpublished (optional): Include unpublished apps (default: true)

Returns: Object containing paginated apps, streams, and pagination metadata

Example (default - first 50 apps):

{
  "apps": [...],
  "streams": [...],
  "pagination": {
    "limit": 50,
    "offset": 0,
    "returned": 50,
    "total_found": 1598,
    "has_more": true,
    "next_offset": 50
  },
  "filters": {
    "name_filter": null,
    "app_id_filter": null,
    "include_unpublished": true
  },
  "summary": {
    "total_apps": 1598,
    "published_apps": 857,
    "private_apps": 741,
    "total_streams": 40,
    "showing": "1-50 of 1598"
  }
}

Example (with name filter):

# Search for apps containing "dashboard"
result = 

---