8/11/2025

So You Want to Connect Claude to Your Dataverse? A Guide to Setting Up a Dataverse MCP Server

Hey everyone, let's talk about something pretty cool you can do if you're working with Microsoft Dataverse & you're also a fan of AI assistants like Anthropic's Claude. Turns out, you can actually connect them. This means you can use natural language in Claude to directly query, update, & interact with your Dataverse tables. It's a HUGE time-saver & a big step up in terms of workflow efficiency.
Honestly, getting it all set up can feel a bit technical at first glance, but it's totally manageable if you take it step-by-step. I've been through the process, & I'm here to walk you through it. We're going to set up what's called a Dataverse MCP Server.
"MCP" stands for Model Context Protocol. Think of it as a special translator that lets large language models (LLMs) like Claude securely talk to external data sources—in this case, your Dataverse environment. By setting this up, Dataverse acts as a server, & Claude becomes the client, able to make requests & get information back in a structured way.

Why Bother Doing This?

Before we dive into the "how," let's quickly cover the "why." Imagine you're building a report or analyzing customer data. Instead of manually navigating through Power Apps or writing complex queries, you could just ask Claude:
  • "Hey, how many active accounts do we have in the system?"
  • "List all the contacts created in the last week."
  • "Create a new task for the 'Contoso' account with the description 'Follow up on Q3 proposal'."
This is incredibly powerful for developers, business analysts, & anyone who needs to interact with Dataverse data without the usual overhead. It streamlines your workflow & lets you leverage the conversational power of AI to manage your business data.
This whole idea of using AI to interact with business systems is a game-changer. It's the same principle behind a lot of modern business automation. For instance, companies are increasingly using AI to handle customer interactions. This is where a platform like Arsturn comes into play. Businesses use Arsturn to build no-code AI chatbots trained on their own data. These bots can then engage with website visitors 24/7, answering questions, providing instant support, & even generating leads. It's all about creating a more efficient & personalized experience, whether it's for your internal team talking to Dataverse or your customers talking to your website.
Okay, let's get into the nitty-gritty of the setup.

The Game Plan: What You'll Need

Here’s a high-level look at what we're going to do. Don't worry, I'll break each of these down.
  1. Handle the Prerequisites: Make sure you have the right environment & tools ready to go.
  2. Create a Dataverse Connection: This is a specific connection that the MCP server will use.
  3. Install the Dataverse MCP Server Local Proxy: This is a small tool that runs on your computer to facilitate the connection.
  4. Grab Your Tenant ID: A unique identifier for your environment.
  5. Configure Claude Desktop: This is the final step where you tell Claude how to find & use your new Dataverse server.
Let's do this.

Step 1: Getting Your Ducks in a Row (Prerequisites)

First things first, you need a couple of things in place before you start.
  • A Power Platform Environment with Dataverse: This is probably obvious, but you need an active Dataverse instance. This won't work with a regular Power Apps environment that doesn't have a Dataverse database.
  • The Right Permissions: You'll need to be an admin or have sufficient permissions in the environment to create connections.
  • .NET 8.0 SDK: The local proxy tool we're going to install is built on .NET. So, you need the SDK (Software Development Kit) installed on your machine. You can get it directly from the Microsoft website or, if you're a command-line fan, you can use PowerShell & run:
    1 winget install Microsoft.DotNet.SDK.8
  • Claude Desktop App: This setup is specifically for the Claude desktop application. You'll need to have it downloaded & installed on your Mac or Windows machine.
Once you've got all that sorted, you're ready for the real fun to begin.

Step 2: Creating the Dataverse Connection

The MCP server needs a dedicated connection to your Dataverse environment to authenticate & interact with it. Here’s how you create one:
  1. Head over to Power Automate: Go to
    1 make.powerautomate.com
    .
  2. Check Your Environment: In the top-right corner, make sure you have the correct Power Platform environment selected. This should be the one that contains the Dataverse data you want to connect to.
  3. Navigate to Connections: On the left-hand navigation pane, find & select "Connections."
  4. Create a New Connection: Click the "+ New connection" button on the command bar.
  5. Find the Dataverse Connector: In the search box, type "Dataverse." You're looking for the one with the green icon, officially called the "Microsoft Dataverse" connector. Select it.
  6. Authenticate: Follow the on-screen prompts to create the connection. You'll likely just need to sign in with your work account.
  7. IMPORTANT - Copy the URL: Once the connection is created, it will appear in your list of connections. Click on it to open its details page. Now, copy the ENTIRE URL from your browser's address bar. It'll look something long & complicated, like
    1 https://make.powerautomate.com/environments/.../connections/...
    . Paste this URL into a notepad or somewhere safe. We're going to need it later.
That's it for this step. You now have the dedicated doorway for your MCP server.

Step 3: Installing the MCP Server Local Proxy

This part sounds intimidating, but it's really just a single command. The "local proxy" is a lightweight, command-line tool that runs in the background on your computer. When Claude wants to talk to Dataverse, it sends the message to this local proxy, which then securely forwards it to your Dataverse environment.
Here’s how to install it:
  1. Open a Terminal or Command Prompt: On Windows, you can open Command Prompt or PowerShell. On a Mac, open Terminal.
  2. Run the Installation Command: Type or paste the following command & hit Enter:

Copyright © Arsturn 2025