What it checks

Three dimensions, column level

Your tables are scored across three dimensions down to each individual column. Checks run automatically against thresholds you define, are fully historized, and can run on a schedule.

Completeness

Detects missing or NULL values per column. Ensures that every field expected to hold data actually does, so downstream consumers can rely on complete records.

Uniqueness

Surfaces duplicates in your primary and business keys. Catches the silent data integrity issues that propagate errors into reports and analytics.

Referential integrity

Identifies foreign keys with no matching parent row. Validates that relationships between tables hold and that joins produce the results you expect.

How it works

From config to results in one click

Describe what to audit, launch the scan, read the results. No infrastructure to manage, no data to move.

01

Describe

List the columns to audit or drop in a DBML file. The app builds one test per check automatically.

02

Scan

One click launches the full scan inside an ephemeral Snowpark Container Services job. Everything runs in your account.

03

Analyze

Results land in a Streamlit interface: global score, table by metric heatmap, worst performing columns, and trend over time.

Why it's different

Built for Snowflake, not bolted on

No SaaS middleman, no data extraction, no new vendor to onboard. Your data stays where it is.

Architecture

100 % in account

No external service, no data movement. Every check runs natively inside your Snowflake environment.

Execution

One click runs

On demand or scheduled, with full run history. Every past result is stored and comparable.

Control

Your thresholds

Set acceptable quality levels per metric and per column. You define what "good enough" means for your data.

Onboarding

Built in demo dataset

Scan something real in minutes, then point the tool at your own tables when you're ready.

Business needs

Data quality and cleansing

Bomzai Data Quality audits any table in your Snowflake account across three dimensions, down to the individual column:

Completeness

Are values present where they should be?

Uniqueness

Are there unexpected duplicates?

Referential integrity

Do relationships between tables hold?

You set the thresholds that define acceptable quality for your own data, run checks on demand or on a schedule, and keep a full history of every result.

Everything runs inside your account: no data moves, no external service.

Issues surface in a Streamlit interface: a global score, a table by metric heatmap, and your worst performing columns. Your teams know exactly what to clean before the data feeds reporting or downstream pipelines.

Ready to score your data?

Get a live demo on your own Snowflake account. See your completeness, uniqueness, and referential integrity scores in minutes.