Live console
Look at the real console before you install anything
This is the operator console StepStitch actually ships, running against a synthetic dataset built by the real pipeline. No signup, no credentials, nothing to deploy. It is read-only: nobody can change what the next visitor sees.
The hosted console is offline right now.
Nothing is lost: it is the same page the project serves at /dashboard, and one command brings it up locally with the same synthetic data.
Six failures, one in every state
A bug does not go straight from reported to fixed, and a console that only shows those two ends is hiding the part you actually work in. The demo puts one failure in each state so the whole path is visible at once.
Fixed and proven
The reproduction failed on the buggy commit and passed on the fix. Both runs happened. That pair is the only thing StepStitch accepts as proof, and it is what fills the memory the console matches new reports against.
Waiting for a test run
Reported, scrubbed and scored. Nobody has run the generated test yet.
Seen before
Structurally close to something already fixed, so the console says where to start.
Confirmed broken
The reproduction failed as expected. The bug is real and reproducible on demand.
Still broken
A fix was tried. The same test still fails, so the verdict says so.
Test needs fixing
The reproduction passed on the broken build, so it never reproduced anything.
Waiting for a test run
0no CI result yet
Seen before
1you fixed this before
Transfer — the server errored after Confirm transfer
1 person affected
You fixed this before — see PR-42
Confirmed broken
1reproduced, awaiting a fix
Checkout — the request was rejected
1 person affected
Fixed and proven
1failed before the fix, passed after
Transfer — the server errored after Review transfer
4 people affected
Four reports of one bug are one card, not four rows. Engineers can flip on technical detail for route templates, status codes and raw payloads.
What is synthetic, and what is not
The failures are invented. Everything done to them is the shipping code.
Made up
A fictional bank, six invented failures, and fake customer details written into the reports on purpose. There is no real person behind any of it, and the account numbers and emails you will see redacted were never real.
Real
The scrubber that removed those details, the replayability score, the generated Playwright test, the red-to-green verdict and the hashed attestation are all produced by the same code a self-hosted install runs. The dataset is regenerated from that pipeline and a test fails if it drifts.