AutoDBTTM

Half your dbt code is dead.We bury it safely.

  • ranked by what it costs
  • proof before every merge
  • your team owns the merge

scanequivalence proofPR you approve

Chat with FoundersSee how it works
dbtBuilt for dbt teams

[ how it works ]

Every cleanup arrives as a PR, with the proof attached.

github.com/loyalty-co/dbt · PR #1842
Open

Merge fct_orders_enriched into fct_orders

zingle-bot opened 2 minutes ago · cleanup/merge-fct-orders → main

+47-132·7 readers verified·95% SQL matchsaves $3,200/mo
Equivalence report100% row match across 12.84M rows
Blast radius17 readers verified · all preserved
Backfill planidempotent · 60s rollback if needed
CI · dbt buildpassed in 4m 12s

$47k

average monthly warehouse spend recovered

per team, in first 90 days

40%

of dbt models are never read after 90 days

sitting in your repo right now, costing money

minutes

from connecting your repo to your first PR

not days. AutoDBT scans and acts immediately

[ find the debt ]

See the debt in your repo. Ranked by what it costs you.

Duplicate models

That's why your numbers don't tie

95% the same. One different filter. That's why your numbers don't tie. AutoDBT finds every near-duplicate and ranks them by cost overlap.

Near-duplicate models

fct_orders_enriched

$3.2k/mo
97% SQL match

fct_orders

$2.8k/mo
97% SQL match
filter scope differs · merge safe

[ ship it safely ]

Every change comes with proof. Your team decides what merges.

Equivalence testing

Row for row, or it doesn't merge

Row-level diff before every merge. Configurable tolerance. If output drifts by even 0.01%, the PR is blocked automatically.

Row-level equivalence

12.84M rows
100%match

Tolerance

0.01%

Scope

mart_orders

Regions

all · passed

Runtime

4m 12s

[ read next ]

See what is hiding in your repo.

run log · Leading consumer tech35% of compute, removed on a cron110 PRs, and the 0.4% two dashboards disagreed on.Read →use caseMulti-ComputeOnce you've moved, every workload on the cheapest engine that fits.Read →essayWhy data is differentRL dies on the reward function. Data was born with one.Read →

Working on something like this? Chat with Founders →