Grain Inspector
Find Out What One Row Actually Represents
A Row Scale, not a schema diagram. Drop candidate grains — Account, Subscription, Invoice, User, Event — onto the scale. A model with mixed grain tips it immediately, before the mixed rows tip a metric downstream.
Model under inspection · fct_billing_events
Scale unbalanced — Open Key Evidence for the primary-key and uniqueness proof behind the tip.
- Inspect Model — load a warehouse table or dbt model
- Test A Candidate Grain — check if a key is truly one row per entity
- Open Key Evidence — distinct counts and row counts, side by side
- Split Mixed Grain — separate rows that don't share one grain
- Confirm Grain — lock the verified grain for downstream use
A Correct Query Can Still Produce The Wrong Metric.
Illuminate The True Grain Of Every Metric.
GrainLux is the AI Semantic Data Reliability Layer between your warehouse and every model, metric, and dashboard built on top of it — grain-aware, metric-native, explainable, and deterministic.