Everything you need to get answers faster

1. Create an account (or open the live demo — no sign-up needed).
2. The Quick-Commerce Orders sample is loaded automatically.
3. Open AI Agents and ask the Advisor for an executive summary.
4. Upload your own CSV from Datasets whenever you’re ready.

• Supported format: CSV (UTF-8) with headers in the first row, up to 25 MB.
• From Excel: File → Save As → CSV UTF-8. From Google Sheets: File → Download → .csv.
• Types are detected automatically: number, date (YYYY-MM-DD), category, text and ID.
• Files are parsed in your browser. With an account they are saved privately to your workspace (row-level security); raw rows are never sent to the AI.

• Cleaner — profiling, quality score, cleaning plan and pandas code.
• SQL — natural language to SQL, executed in-browser, explained clause by clause.
• Viz — chart recommendations and Power BI-ready dashboard layouts.
• Marketing Analytics — segmentation, channel ROAS/CAC, RFM and campaign plays.
• Advisor — executive summary, insights, top 3 recommendations and KPIs.

• Your active dataset is always the table data.
• Use standard SELECT, WHERE, GROUP BY, ORDER BY, LIMIT and aggregates (SUM, AVG, COUNT, MIN, MAX, ROUND).
• Month grouping: SUBSTRING(order_date, 1, 7).
• Run with ⌘/Ctrl + Enter; results can be viewed as a table or chart and exported to CSV.

• Daily series per metric, compared against a rolling 14-day baseline (7–30 configurable).
• Robust z-score (median/MAD, log space). Flag when |z| ≥ threshold (default 3σ); critical when |z| ≥ 5.
• Each alert includes the top contributing value of your main dimension.

• With a Google Gemini key configured (server env GEMINI_API_KEY or your own key in Settings), agents answer generatively.
• Without a key, the built-in DataOps engine answers from real statistics computed on your data.
• Only a compact profile (column names, types, counts, averages, top categories) is sent to the model.

DataOps was designed and built as an end-to-end analytics product: a Next.js front end, an in-browser analysis engine, and an LLM layer for natural-language insights.