Talk to your data through natural language. No SQL or technical background required for data query, analysis, and visualization.
User question → Parse intent → Generate SQL → Execute query → Analyze results → Visualize → Output conclusions
Chinese questions auto-converted to SQL:
| User Question | SQL |
|---|---|
| --------------- | ----- |
| "Sales by region last month?" | SELECT region, SUM(amount) FROM sales WHERE month='2026-04' GROUP BY region |
| "Which product has the highest return rate?" | SELECT product, COUNT(*) FROM orders WHERE status='returned' GROUP BY product ORDER BY 2 DESC LIMIT 1 |
| "Compare user growth vs same period last year" | SELECT DATE_TRUNC('month', created_at), COUNT(*) FROM users WHERE created_at >= NOW() - INTERVAL '1 year' GROUP BY 1 ORDER BY 1 |
Results output in two layers:
| Mode | Description |
|---|---|
| ------ | ------------- |
| Refine | "Only show East China" → append filter |
| Switch dimension | "Group by month" → re-aggregate |
| Root cause | "Why did it drop?" → drill down |
| Compare | "vs last quarter?" → time comparison |
Supports Excel (.xlsx/.xls), JSON/JSONL, CSV file reading. See references/data-sources.md.
MySQL / PostgreSQL / Snowflake / SQLite / BigQuery / Redshift. See references/data-sources.md.
Inline chat output:
📊 Results: {title}
─────────────────────────────
{metrics table}
📈 Trend:
{ASCII trend bars}
📋 Analysis:
1. ...
For charts, auto-generate HTML page → write to {domain}_chart.html → report path.
共 1 个版本