Chart builderLive

When the question isn't on a screen yet.

Drag a measure. Drag a dimension. Pick a chart type. Every model in the semantic layer is queryable right here, so the answer is one drag away even when nobody built the dashboard for it.

MetricsBuilderModelUnified Customer Activity
ModelUnified Customer ActivityMetrics6 selectedRows2 selectedColumns1 selected
Total Purchase Revenue > 07 rows
Region / CityAdd to CartCheckout CVRSession → CVRRevenue ₹
Punjab2.00100.000.921,999
Ludhiana2.00100.000.921,999
West Bengal2.0050.001.921,999
Kolkata2.0050.001.921,999
Delhi Division5.00150.003.1816,193
New Delhi2.00100.002.6314,394
Haryana3.0066.678.703,979
Gurugram3.0066.678.703,979
Tamil Nadu1.00100.0014.291,999
Madurai1.00100.0014.291,999
Total14.00483.3429.3628,168

The real builder, recoloured. Flip the chart type and the same query redraws.

01

Four moves. One canvas.

Every model in the semantic layer is one drag from a chart, including the ones we never built a screen for.

Pick a model

Sales, marketing, cohorts, fulfilment, quick commerce, B2B, planning. Everything modelled in the semantic layer is queryable here, not just the screens we pre-built.

SalesMarketingCohortsFulfilmentQuick commerceB2BPlanning

Drag measures and dimensions

Compose a pivot, layer filters, add a comparison window. The semantic layer writes and runs the query underneath, so there is no SQL to learn.

Switch how you see it

The same query as a pivot, a line, a bar or an area. One toggle, no re-querying.

Save and share

Group related charts into a dashboard, set a refresh cadence, pin it to the team board. Per-role views open straight in Slack, and everyone reads the same numbers because everyone reads the same model.

Finance viewGrowth viewOps viewone model
02

Pick how you see it.

One query, every shape. The format is a toggle, never a rebuild.

PivotLineBarAreaTarget vs actualData table
03

No SQL to write. None to read.

You compose with measures and dimensions. The semantic layer turns that into a governed query, runs it, and hands back the number. The query stays under the hood.

You speak measures. The layer does the rest.

Drag the fields you care about. The semantic layer resolves them to agreed definitions, writes the query and runs it, so there are no tables to join and no SQL to learn.

  • Compose with governed measures and dimensions
  • The semantic layer writes and runs the query for you
  • One definition per number, shared across every team
You pick
Checkout CVRSession → CVRRevenue ₹Region / City
Semantic layer writes and runs the query
You get
One governed number, defined once and read the same by every team.