Skip to content

quickstart.md

Latest commit

 

History

History
106 lines (76 loc) · 2.8 KB

quickstart.md

File metadata and controls

106 lines (76 loc) · 2.8 KB

Quick Start — MCP Analytics

Get running in under a minute.

Step 1 — Install in Your AI Client

Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "mcp-analytics": {
      "command": "npx",
      "args": ["-y", "mcp-remote@latest", "https://api.mcpanalytics.ai/auth0"]
    }
  }
}

Restart Claude Desktop completely.

Cursor

Add to your Cursor MCP settings (Settings → MCP):

{
  "mcpServers": {
    "mcp-analytics": {
      "command": "npx",
      "args": ["-y", "mcp-remote@latest", "https://api.mcpanalytics.ai/auth0"]
    }
  }
}

VS Code (Continue extension)

Add to ~/.continue/config.json:

{
  "mcpServers": {
    "mcp-analytics": {
      "command": "npx",
      "args": ["-y", "mcp-remote@latest", "https://api.mcpanalytics.ai/auth0"]
    }
  }
}

Step 2 — Sign In

On first use, your AI client will open a browser window for OAuth sign-in. Create a free account at mcpanalytics.ai or sign in if you already have one. This is a one-time setup — your client stores the token and refreshes it automatically.

Step 3 — Start Analyzing

Once connected, just ask your AI assistant:

"I have a CSV of monthly sales data. Can you forecast next quarter?"

"Analyze this Shopify orders export and tell me what's driving revenue"

"Run a t-test on these two groups and tell me if the difference is significant"

The AI will:

  1. Discover the right analysis tool for your question
  2. Ask you to upload your data (or connect a data source)
  3. Run the analysis
  4. Return an interactive HTML report with charts and insights

Free Tier

The free tier includes 25 analyses per month — no credit card required. See pricing for paid plans.

Try a Live Demo

Before connecting, you can try the demo or browse sample reports to see what the output looks like.

Troubleshooting

MCP Analytics doesn't appear in my tools list

  • Restart your AI client completely (not just reload)
  • Verify your config file is valid JSON
  • Check your client's MCP logs for connection errors

Authentication fails

  • Make sure you have a free account at mcpanalytics.ai
  • Try running npx -y mcp-remote@latest https://api.mcpanalytics.ai/auth0 directly in a terminal to test the connection

Analysis fails

  • Check that your CSV has column headers in the first row
  • Try describing your data more specifically so the AI picks the right tool
  • Contact support@mcpanalytics.ai with the error message

Full docs → | Support →