ConvergeQL by Measured Studios

Data models that converge before they ship.

Turn a BigQuery or Snowflake model change into a reviewable GitHub change with explicit model, data, cost, lineage, and governance evidence.

Inspect an example changeRead what converged means

Public product preview. BigQuery and Snowflake adapters remain at foundation and have not connected to a customer environment.

Synthetic ChangeRun

Optimize customer revenue

Converged
Protected grain
customer_id + month
Delivery boundary
Human-approved Git change
Platform
Illustrative BigQuery profile
Policy snapshot
model-review-v3
  1. Candidate 01Output regressionRevenue totals changed by 0.7%
  2. Candidate 02Cost policyEstimated processing above declared limit
  3. Candidate 03Required gates9 of 9 declared checks passed

Illustrative workflow only. This is not customer, warehouse, savings, or production evidence.

The review problem

SQL generation is getting cheaper. Trustworthy model change is not.

Senior data engineers still resolve source authority, grain, relationships, regressions, cost, governance, and downstream impact before a shared model can ship.

ConvergeQL makes that discipline visible as a bounded run with an editable specification, rejected candidates, deterministic gates, and evidence a reviewer can inspect.

One governed change

The model decision comes before the diff.

Chat may start or steer a run, but the durable control surface is the model specification and the evidence attached to it.

01

Model specification

Intent, grain, keys, protected outputs, and constraints.

02

Candidate history

What was rejected, what changed, and which evidence caused the decision.

03

Provider evidence

Platform-specific validity, cost, usage, and lineage observations.

04

Git delivery

A reviewable change after the declared gates pass and a human approves.

Convergence, made literal

Probabilistic search. Deterministic acceptance.

A candidate converges only when every mandatory predicate in the recorded policy passes against the same context snapshot.

01Intent requirementsPass
02Schema contractPass
03Declared grainPass
04Data regressionPass
05Query validityPass
06Cost policyPass
07Lineage impactPass
08Governance policyPass
09Review materialPass

Synthetic gate display. A future run must attach validator identity, evidence, limitations, and context version.

Data model flow

From modeling intent to reviewable change.

Explore the complete workflow
  1. 01Discover
  2. 02Specify
  3. 03Model
  4. 04Validate
  5. 05Converge
  6. 06Review
  7. 07Ship
  8. 08Observe

Independent platform profiles

Same product contract. Different evidence.

BigQuery dry-run evidence and Snowflake EXPLAIN evidence keep their own methods, units, timing, and limitations.

Google BigQuery

foundation

Dry-run and job-evidence contracts are defined; no project connection exists.

Query validity
contract defined
Scan estimate
partial by platform
Observed usage
contract defined
Lineage
partial by platform

Snowflake

foundation

EXPLAIN and query-evidence contracts are defined; no account connection exists.

Query validity
contract defined
Scan upper bound
partial by platform
Observed usage
contract defined
Lineage
partial by platform
Inspect platform evidence boundaries

Design partner preview

Help shape the first governed model-change workflow.

Bring one existing dbt-managed model, its declared invariants, and a review problem worth solving.

Discuss a design partnership