Turn Your AI Assistant Into an Azure Data Explorer Expert

Your AI assistant knows how to write code, debug issues, and analyze patterns—but it can't directly access your Azure Data Explorer databases. Until now.

This MCP server bridges that gap, giving Claude or any MCP-compatible AI assistant direct querying capabilities against your ADX clusters. No more copying query results back and forth or explaining your data schema repeatedly.

Stop the Context Switching Dance

You know the routine: ask your AI assistant to help analyze some data, switch to ADX to run the query, copy results back, paste them into chat, ask follow-up questions, repeat. It's tedious and breaks your flow.

With this MCP server, your AI assistant can:

• Execute KQL queries directly against your ADX databases
• Discover table schemas and relationships automatically
• Sample data to understand structure and patterns
• Iterate on queries based on results without manual intervention

Real Productivity Gains

Exploratory Data Analysis: Instead of manually running take 100 queries to understand your data, your AI assistant can explore table schemas, sample representative data, and build understanding of your dataset structure autonomously.

Complex Query Development: Start with "show me the top 10 errors from last week" and let your AI assistant build increasingly sophisticated KQL queries based on the actual data it finds, refining the analysis in real-time.

Incident Response: During outages, your AI assistant can query telemetry data directly, correlate events across tables, and identify patterns—all while you focus on the fix rather than data wrangling.

Built for Production Workloads

The server handles Azure authentication properly, supporting both traditional service principals and Azure Workload Identity for AKS deployments. You can run it locally during development or deploy it containerized in your production environment.

# Quick start - works with your existing Azure CLI login
uv run src/adx_mcp_server/main.py

The tool discovery is configurable, so you can expose only the capabilities you need without cluttering your AI assistant's context window.

Integration That Just Works

Configure it once in Claude Desktop and forget about it:

{
  "mcpServers": {
    "adx": {
      "command": "uv",
      "args": ["--directory", "/path/to/adx-mcp-server", "run", "src/adx_mcp_server/main.py"],
      "env": {
        "ADX_CLUSTER_URL": "https://yourcluster.region.kusto.windows.net",
        "ADX_DATABASE": "your_database"
      }
    }
  }
}

Your AI assistant now has persistent access to your data infrastructure. No API keys to manage, no additional authentication flows—it uses your existing Azure credentials.

Perfect for Azure-Heavy Workflows

If you're already invested in the Azure ecosystem with ADX for telemetry, logs, or analytics, this server makes your AI assistant a first-class citizen in your data workflows. It understands KQL syntax, can navigate your table relationships, and works with the data formats you're already using.

The Docker support means it fits naturally into your existing container orchestration, and the comprehensive test suite ensures it's reliable enough for production use.

Ready to eliminate the friction between your AI assistant and your data? Your next data exploration session is about to get a lot more conversational.