Daggy integrates directly with AI tools like Claude desktop to allow you to perform advanced data analytics and machine learning modeling. You can think about it as an application for your AI assistant. Your assistant can orchestrate tasks and send them to Daggy to perform analysis on your data.

Daggy works via the Model Context Protocol (MCP), which allows AI systems to interact with external tools and data sources, extending their functionality. The setup instructions below show you how to connect your AI assistant to Daggy via MCP.

Quick Setup

Prerequisites

Ensure you have:

• Node.js installed on your system
• Access to Claude Desktop, Copilot through Visual Studio Code, or another MCP client like Codename Goose or Raycast.

Installing & Connecting to MCP

Daggy uses the Model Context Protocol (MCP) to connect AI systems with external tools. Here's how to get started:

Integrate with Claude Desktop

To connect Daggy with Claude Desktop:

1. Open Claude Desktop
2. Navigate to Settings → Developer → Edit Config
3. Edit Config will open claude_desktop_config.json. Replace the file with the following:

{
  "mcpServers": {
    "daggy_mcp_server": {
      "command": "npx",
      "args": ["mcp-remote", "https://mcp.daggydata.com/sse"]
    }
  }
}

4. Save the file and restart Claude Desktop

Integrate with Copilot through Visual Studio Code

To connect Daggy with Copilot through Visual Studio Code:

1. Open Visual Studio Code
2. Open the Command Palette by navigating View → Command Palette or by pressing Ctrl+Shift + P (Windows/Linux) or Cmd + Shift + P (macOS).
3. Type MCP in the search bar, and select MCP: Add Server
4. Use Command (stdio) and enter the following command: npx mcp-remote https://mcp.daggydata.com/sse
5. Then enter daggy-mcp-server, which will be the server's name. You can either select Global or Workspace.
6. An mcp.json file will pop up, that looks like

{
	"servers": {
		"daggy-mcp-server": {
			"type": "stdio",
			"command": "npx",
			"args": [
				"mcp-remote",
				"https://mcp_dev.daggydata.com/sse"
			]
		}
	},
	"inputs": []
}

7. Go back to the Command Palette and type MCP, then click on List Servers. Find daggy-mcp-server and click it. Then click Start Server.
8. Now open Copilot/Chat. Make sure you are on Agent mode, rather than Chat or Edit mode.
9. Now ask Copilot something like "What datasets do I have available in Daggy?" and it will prompt a login.

Integrate with Codename Goose

To add Daggy as an extension in Codename Goose:

1. Install Codename Goose if you haven't already

2. Follow the extension installation instructions

3. Configure the Daggy extension with these settings:

   • Extension Name: Daggy
   • Command: npx mcp-remote https://mcp.daggydata.com/sse
   • Description: Performs advanced statistical analysis and machine learning on data
   • Timeout: 300

4. Save your configuration and toggle the Daggy extension on in Goose.

Testing Connection

Once connected, you will be prompted to login with your Daggy username and password. After logging in, your AI assistant will be able to leverage Daggy's powerful data analysis capabilities. Test your connection by asking your AI assistant something like:

• "Write a dataset to Daggy that shows the relationship between coffee consumption and productivity. Then tell me how productive I'd be if I drank 10 cups of coffee per day."
• "What datasets do I have available in Daggy?"

Managing Data

Uploading Data

There are multiple ways to get your data into Daggy for analysis:

1. Via Daggy Web Interface (CSV Only)

CSV Format Requirements

Ensure your CSV files follow these guidelines:

• Header Row: Include exactly one header row at the top with column names
• No Metadata Rows: Don't include title rows, notes, or other content above the headers
• Clean Data: Remove any blank rows or extraneous columns

Upload Steps

1. Go to app.daggydata.com/data
2. Drag and drop your CSV file or click to browse
3. Add a helpful description (optional)
   • Example: "Daily S&P 500 (SPY) data from January 2010 to December 2025"
4. Your data will be immediately available to your AI assistant

2. Via AI Assistant (Claude Desktop, Goose, etc.)

Structured Data (CSV) Drag and drop a CSV file into your AI assistant conversation, then ask your assistant to upload it to Daggy.

Example: "I've uploaded a CSV with sales data. Can you upload this to Daggy and analyze the relationship between sales and price?"

Unstructured Data (PDFs, JSON, Web, etc.) Your AI assistant can extract and structure data from various formats:

• PDF Data Extraction - Extract tables and data from documents
• Web Scraping - Pull data from websites and APIs
• JSON/XML Processing - Parse structured data files
• Other MCP Servers - Connect to other MCP servers and pull data
• Generated Datasets - Create synthetic data for testing

Example prompts:

• "Extract the quarterly sales from this PDF and upload to Daggy"
• "Pull my Square sales data from the last 30 days into Daggy"
• "Scrape this website's stock price table and write it to Daggy"

Confirming Success Ask your AI assistant a question about your data - it will answer through Daggy, confirming the upload worked.

Deleting Data

To remove datasets, simply ask your AI assistant to delete them by name.

AI Agent Commands

AI agents are extremely adept at fulfilling your requests. Although these agents can figure out what Daggy tools to use to answer your query, the following commands can help you get started.

Data related commands:

• What data do I have in Daggy?
• Write data to daggy (after you've dragged data into the AI assistant or had it generate data some other way)
• Delete dataset X from daggy

Model related commands

• What models do I have in daggy?
• Build a model that predicts y from columns a, b, c in dataset K
• Run a scenario with model X from daggy
• Delete model X from daggy

These are just basic commands. Check out our blog for more examples.

You're Ready!

You can now ask your AI assistant to perform advanced analytics on your data, from Monte Carlo simulations to predictive modeling.

See examples in our blog.