Skip to main content

​
Overview

The AI Engine lets you ask questions about your data in plain English. Plumi’s AI understands your database schema and generates accurate SQL queries automatically.

​
How It Works

  1. Schema Understanding: The AI learns your database structure, including tables, columns, relationships, and data types.
  2. Natural Language Processing: Type your question in plain English, and the AI interprets your intent.
  3. SQL Generation: The AI generates optimized SQL that answers your question.
  4. Execution & Results: Execute the query and view results instantly.

​
Supported AI Models

Plumi supports multiple AI providers:
ProviderModels
AnthropicClaude 3.5 Sonnet, Claude 3 Opus
OpenAIGPT-4, GPT-4 Turbo
GoogleGemini Pro
MistralMistral Large
Configure your preferred model in Settings > AI Models.

​
Example Queries

​
Simple Questions

“How many orders did we get last month?” 
SELECT COUNT(*) as order_count
FROM orders
WHERE order_date >= DATE_TRUNC('month', CURRENT_DATE - INTERVAL '1 month')
  AND order_date < DATE_TRUNC('month', CURRENT_DATE);

​
Complex Analysis

“Show me the top 5 products by revenue, broken down by month for Q4 2024” 
SELECT
  DATE_TRUNC('month', o.order_date) as month,
  p.product_name,
  SUM(oi.quantity * oi.unit_price) as revenue
FROM order_items oi
JOIN orders o ON oi.order_id = o.id
JOIN products p ON oi.product_id = p.id
WHERE o.order_date >= '2024-10-01' AND o.order_date < '2025-01-01'
GROUP BY DATE_TRUNC('month', o.order_date), p.product_name
ORDER BY month, revenue DESC;

​
Best Practices

Be specific: Instead of “show me sales”, try “show me total sales by region for the last 30 days”
Reference tables: If you know the table name, mention it: “from the orders table, show me…”
Review generated SQL: Always review the generated query before running it on production data

​
Configuring AI Models

​
Adding API Keys

  1. Go to Settings > API Keys
  2. Add your API key for your preferred provider
  3. Select the model in Settings > AI Models

​
Model Selection

Choose models based on your needs:
  • Claude: Best for complex analytical queries
  • GPT-4: Great general-purpose model
  • Gemini: Fast and cost-effective
  • Mistral: Good balance of speed and quality