Using AI Features

Use natural language to generate SQL - describe what you want, get the query.

Setup

  1. Click the gear icon to open Settings
  2. Enter your API key from your LLM provider
  3. Select your provider (OpenAI, Anthropic, Google, etc.)
  4. Choose a model
  5. Click Test Connection to verify it works
  6. Click Save

Your API key is stored in your browser only. It's sent directly to the provider's API - it never passes through ExploreMyData's servers.

AI Intent in Operation Panels

Many operation panels have an AI toggle button. When enabled, you can describe what you want in plain English instead of writing SQL:

  • Filter: "Show orders from January where revenue is over 1000"
  • Add Column: "Create a column that calculates profit margin as (revenue - cost) / revenue"
  • Update Values: "Set status to 'overdue' where ship_date is null and order_date is more than 7 days ago"
  • Math: "Calculate the running average of revenue by category"

The AI generates a SQL expression. Review it, edit if needed, then apply.

Pipeline Natural Language

Type a description in the pipeline input at the bottom of the screen. The AI generates a full SQL step.

Examples:

  • "Group by category and show the average revenue"
  • "Remove duplicate rows based on email"
  • "Add a column with the day of week from order_date"

Tips for Better Results

  • Be specific: "Filter to orders from Q1 2024" works better than "show recent orders"
  • Use column names: The AI knows your column names, so reference them directly
  • Review the SQL: Always check the generated SQL before applying
  • Iterate: If the result isn't right, edit the SQL or try rephrasing

Supported Providers

ExploreMyData supports multiple LLM providers. Configure your preferred one in Settings. All API calls go directly from your browser to the provider.