Owned local execution · 15 October 2026

Inspect a real run of the testing engines.

Follow an invoice fault from an independent requirement to the executed check and its recorded result. This run uses Cue April’s actual testing engines on fictional data.

Executed locally, with no hosted runner, customer data or AI calls. It demonstrates these engines and fixtures; it does not qualify a hosted evaluation.

Tested product revision
ba079ad32f5d8db588ac3314ee31ec8d5809d75a
Independent specification
demo-intent-v1
Recorded runtime
Node v20.19.4; fast-check 4.3.0
Recorded execution

Compare each engine on the same faults.

The corrected code passes the included checks. The delivery fault charges £4.99 at the exact £50 threshold. The access fault lets one fictional tenant read another tenant’s invoice. Passing unrelated checks does not clear either fault.

Actual recorded outcomes, including deliberately failing checks
Code versionCheck and toolOutcomeEvidence
Corrected codesample-unit
sample-node-unittest@1.0.0
PassedRead result ↗
Corrected codesample-api
v20.19.4
PassedRead result ↗
Corrected codesample-property
tenhaw-greedy-covering-array-v1
PassedRead result ↗
Delivery boundary faultsample-unit
sample-node-unittest@1.0.0
FailedRead result ↗
Delivery boundary faultsample-api
v20.19.4
FailedRead result ↗
Delivery boundary faultsample-property
tenhaw-greedy-covering-array-v1
FailedRead result ↗
Invoice access faultsample-unit
sample-node-unittest@1.0.0
FailedRead result ↗
Invoice access faultsample-api
v20.19.4
FailedRead result ↗
Invoice access faultsample-property
tenhaw-greedy-covering-array-v1
PassedRead result ↗
Corrected codeshipping-law
fast-check@4.3.0
PassedRead result ↗
Corrected codetenant-law
fast-check@4.3.0
PassedRead result ↗
Delivery boundary faultshipping-law
fast-check@4.3.0
PassedRead result ↗
Delivery boundary faulttenant-law
fast-check@4.3.0
PassedRead result ↗
Invoice access faultshipping-law
fast-check@4.3.0
PassedRead result ↗
Invoice access faulttenant-law
fast-check@4.3.0
FailedRead result ↗

Keep exact boundaries alongside generated inputs.

The API check expected a £50 invoice and observed £54.99 on the faulty version. The access check expected HTTP 403 and observed 200. Both expectations came from the independent sample specification.

The additional fast-check shipping task used seed 1701 and 100 random examples. It passed while the exact £50 fault remained. The explicit boundary checks caught that fault. A separate tenant property found the access fault and shrank its generated input to [0], with replay path 0:0.

The finite generated-input engine uses tenhaw-greedy-covering-array-v1. It is distinct from the separately recorded fast-check tasks. Neither result establishes that every possible input was tested.

Run the worked fast-check example

Inspect the evidence and reproduce the fixture.

The published JSON contains selected outcomes, tool identity and SHA256 checksums. Private working paths and restricted raw logs are omitted. The downloadable source reproduces the owned fixture and its checks directly; it does not distribute the private Cue April product.

This run did not execute mutation, browser or Gherkin tasks. The Stryker guide and Cucumber guide include separate local educational executions with their own versions and recorded results.

Bring your own requirements to an evaluation.

Your repository needs its own supported setup, qualified runner, scope and spending limit. Your team approves the plan and decides what ships.

Compare ways to start

Your next release starts here

Keep building.
Put testing to work.

Give Cue April your next testing challenge. Let your AI quality lead coordinate the checks, so your team can focus on what comes next.

Create account

Create your account, connect a repository and review your proposed setup before approving a run.