1. Create an API Key

<aside>

You’ll need this for authorization to your Alkemi account.

First, within DataLab, open your account menu in the upper right and select API Keys.

Select the + API Key button to add a new key. Give your key a name. Select Create Key.

And click the clipboard icon to copy it to your clipboard.

</aside>

2. Add Alkemi’s MCP server to your Claude Desktop settings file

<aside>

Open Claud Desktop and go to your settings. Select Developer and then Edit Config.

This will point you to the location of your claude_desktop_config.json file. Open this up and add the following pasting your API Key to replace the text <enter your API key here>

{
    "mcpServers": {
        "alkemi": {
            "command": "npx",
            "args": [
                "mcp-remote",
                "<https://api.alkemi.cloud/mcp>",
                "--header",
                "Authorization: Bearer <enter your API key here>"
            ]
        }
    }
}

Save the config file.

</aside>

3. Chat with Alkemi in Claude!

<aside>

Close and re-open Claude Desktop so it reloads the config file. Open a new chat and click the Search and Tools button to ensure Alkemi is now available for you to invoke.

Ask Claude to run a query that calls your data product. In the example below we have a data product that contains Amazon listing information and we’re asking it to get average prices grouped by sub categories from this data product.

As Claude uses Alkemi’s MCP tools it will ask for permission. It will first ask for access to look for the proper Data Product to query.

Then it will ask for permission to use the tool to query that Data Product

Alkemi will then interpret your query sent from Claude, run the SQL, and send the results back to Claude for interpretation.

That’s it! Now you can access your Data Products and query them from Claude!

</aside>