DataX

Find the queries quietly running up your bill.

Trace warehouse and pipeline cost to the query, the dashboard, the dbt model, and the team that ran it, across Snowflake, Databricks, BigQuery, Fabric, and Synapse.

Book a Demo Explore the platform

Query Attribution preview

Video coming soon

Query Attribution

Every dollar of warehouse spend attributed to the query, pipeline, dashboard, and team that ran it.

Read the docs

Your data warehouse is a blank check.

Warehouse and pipeline cost scales with how people use it, and it rarely maps back to a team or a product. A single query can scan hundreds of gigabytes and cost more than a server. Run it on a schedule and it compounds, quietly, on someone else's budget.Stop waiting for the monthly bill to see who burned the budget.

Before DataX

A blank check

AttributionLumped into one total
WasteHidden in the bill
TimingSeen on the monthly bill
AccessBroad and risky
ActionManual investigation

With DataX

Attributed and controlled

AttributionBy query, dbt model, and team
WasteFull scans and spillage surfaced
TimingDetected as patterns run
AccessRead-only, metadata only
ActionApproved, reversible, audited

Real finding

We find the leaks billing dashboards miss.

This is a real DataX finding. Not a mock. Not an illustration.

Read the docs

The $117 full scan

One query, no partition pruning, scanning everything to return a little.

The pattern billing never shows

The same shape of query run 77 times a month, invisible in a total.

The high-frequency amplifier

A cheap query on a tight schedule that adds up to real money.

warehouse · prod_analyticsFull scan

SELECT user_id, event, ts
FROM   events
WHERE  event = 'click'
-- no partition filter on event_date

Bytes scanned

334.6 GB

Runs / month

Avg cost

$1.52

Total

$117.16

Suggested: filter on the event_date partition

Warehouse intelligence

Warehouse cost intelligence, not just dashboards.

Read the docs

QueryTeamCost

events · full scananalytics$117.16

dbt · fct_ordersdata-eng$42.80

dash · Revenuefinance$18.40

Query attribution

Every query tied to a user, role, dashboard, dbt model, or job. Spend down to the SQL.

×127Full scan

×88Fan-out join

×64Missing prune

Rewrite suggested

Pattern detection

Cost-amplifying patterns caught and explained, with a rewrite suggested.

Projected close

$128k

regression · +38%

Flagged before month close

Predictive signals

Unit-cost regressions surfaced before monthly close, not in the post-mortem.

Automation actions

Resize warehouse M to Sapplied

Auto-suspend idle 5mapplied

Add partition prunepending

Reversible · audited

Safe automation

Partition, cluster, and right-size fixes applied inside policy. Reversible and audited.

Request a free AI cost audit

DataX doesn't profit from your inefficiency.

DataX prices on the structural drivers of your data platform cost, not on billing noise.

What affects pricing

• Platforms connected
• Seats
• Retention window
• Automation scope
• Support tier

What doesn't

• Your total spend
• How much you waste
• How many queries you run
• Warehouse size
• How much we save you

Cost per unit of work

-54% / 6 mo

improving

You pay for your platform, not your waste.

We don't profit from your inefficiency. What you pay reflects your platform, not your problems.

As your platform becomes more efficient, your cost per unit of work improves without being penalised for growth.

Automation without losing control.

Read the docs

DataX applies approved optimisations within the policies you define.

The automation model is about DataX behaviour, not your pipelines.

Controls

• Approval gates
• Scoped permissions
• Reversible actions
• Full audit trail

Automation modes

• Off
• Recommend only
• Approve then apply
• Auto within policy

Automation policy

Off

Recommend only

Approve then apply

Auto within policy

Audit eventlogged

Detected Warehouse oversizing (Snowflake)
Action Resized Medium to Small
Approved Auto-Apply · Safe Optimisation

Built for least privilege.

Read the docs

Connect platforms using read-only access by default. Enable automation only when you are ready: scoped, auditable, reversible.

Connection model

1. Read-only role
2. Metadata only, never your data
3. Scoped to what you approve
4. Every action logged
5. Revoke any time

Read-only means read-only. DataX ingests metadata, query logs, and billing telemetry. It never touches your underlying data, workload code, or runtime configuration unless you explicitly grant automation permissions.

Connection · SnowflakeRead-only

Query historygranted

Warehouse meteringgranted

Billing telemetrygranted

Table contentsnever

Automationopt-in · off

Revoke any time