The first green run
One run. Passing. With proof.
Not a dashboard full of charts, not a backlog of drafted cases. A single automated test that ran to the end, passed, and left a recording you can replay. Until that exists, nothing else in a QA tool is real.
It is also the only milestone we measure as activation — so the thing we ask you to reach is the same thing we judge ourselves on.
Start your first run →What counts as green
A run that finished — it was claimed by a worker and reached a verdict rather than sitting in a queue. A run that passed — every assertion held. And a run that left evidence — a recording of what the browser did, kept with the result.
That last part is the one teams skip and regret. A green tick with nothing behind it cannot be reviewed, argued with, or handed to a developer.
Three ways in. Same destination.
You have requirements
Start from the document
A user story, an acceptance criteria list, a half-finished Confluence page — messy is fine.
- 01Create a project and set the base URL of the app under test.
- 02Upload the document. QAMind drafts test cases from it.
- 03Read them and approve the ones that are right. You are the reviewer, not the typist.
- 04Generate a script for an approved case, then run it.
You have the app
Start from a recording
Install the Chrome recorder (v2.5.30), click through the journey once, and stop.
- 01Record the path a person actually takes — typing, keystrokes, navigation and all.
- 02The recording becomes ordered steps with locators, not a video nobody can replay.
- 03Review the steps, fix anything the page made ambiguous.
- 04Generate the script and run it.
You have neither
Start from one case
The fastest honest start when the app is new and nothing is written down.
- 01Write a single test case — the one path that must never break.
- 02Generate a script for it.
- 03Run it, watch it go green, and let that be the floor you never fall below.
What never leaves your machine
The model that drafts your tests runs on hardware you control. The browser that executes them runs on the tester’s own machine through the Local Runner. The recordings stay with you.
So the thing under test — your application, its data, its DOM, the credentials used to reach it — is never handed to a third-party model. That is not a setting you switch on. It is how the product is built.
Then the second run
Once one test is green, the rest is repetition you no longer do by hand: group cases into a suite, schedule it, and let failures open bugs with the evidence already attached. The first run is the hard one. It is also the only one you have to do today.
Start your first run →