Comparison
Stripe Sigma, Data Pipeline and exports: queryable data, not a detector
Sigma-class tooling is genuinely powerful: SQL over your own Stripe data answers one-off questions beautifully. Everything residual is product work — authoring joins, tolerances, race handling, scheduled execution, alert routing, recovery workflow, re-authoring on schema drift. Hub page for our named native-tooling comparisons.
What raw access gives you
First-party fidelity; ad-hoc forensics; no third party needed for questions you remember to ask on schedule.
What remains yours forever
The detector logic itself, operationalised: tolerances, ordering races, zero-decimal handling, standalone-fee reasoning, backfill pagination, alert thresholds, recovery execution with idempotency.
Complement pattern
Sigma answers ad-hoc questions; a detector owns the queue. Teams running both report neither replaces the other.
Common questions
Do you replace Sigma?
No — different jobs. Our vendor pages cover Sigma and siblings specifically.