AI for QA teams: regression pack automation
Your judgement. mini’s repetition.
mini writes, maintains and runs every regression test. Your team makes the calls.
Regression suitev4.14.046 passed1 failed1 warning
- CriticalThe promo code takes 10% off the subtotalApply a Promo Code · AndroidFixed · mini checks it on the next runAuto-resolved by passing run
- WarningEvery dish on the menu shows a photoBrowse the Menu · ChromeSeasonal dishes have no photo yet, on purpose.Submit feedbackNot a bug
Seasonal dishes have no photo yet, on purpose.
- CriticalThe subtotal matches the items in the cartManage Cart Contents · iOSQA-482
mini takes the repetition.
mini writes the suite.
Scenarios and acceptance criteria, from your app, repo, specs and test cases.
It keeps the suite current.
Adds, edits and retires tests, and lists every change. Revert any.
No selectors to break.
mini reads the screen, so a moved button doesn’t turn the suite red.
The flows scripts can’t reach.
Up to three phones in one test, email codes, deep links, mocked GPS.
Will mini replace us?
No, but the work changes. The rerunning goes to mini; the judgement stays with you.
mini takes
- Writing the regression tests and updating them as the app changes
- Running the whole suite on every release, on iOS, Android and web
- The recording, failed criterion and steps to reproduce for every failure
- Flagging what no criterion asked about, as suggestions you keep or dismiss
Your team keeps
- What counts as a bug, what goes to a ticket and what ships
- Exploratory testing and the new features nobody has tested yet
- The criteria mini tests against, edited whenever you want
Teams that ship with mini.
“We've been working with Minitap for a long time on our mobile apps. They are exceptionally flexible, always ready to adapt their product to our workflow, and their solutions help us verify our apps reliably.”
Mikalai DzmitryieuSenior Software Engineer, Memrise- 2 days → 1 hourFull regression
- 48 hoursTo full coverage
From a pilot customer.
Helmit
Memrise
Nanogram
Impakt
Trippin
MindStash
Clutch
What happens when mini gets it wrong?
Mark the issue Not a bug and tell mini why in a sentence. The issue stops counting against the app, and mini puts your note to use on the next run, so it doesn’t raise the same false alarm on the next run. You can also edit the criterion or mini’s notes about your app.
How do we know a pass is real?
Each acceptance criterion carries its evidence: what mini saw, a screenshot, the moment in the recording and how confident mini is in the grade. If mini can’t grade a test, because the build broke or sign-in failed, it’s marked Unprocessable, never passed.
How long does a full regression take?
Tests run in parallel across your devices and browsers, so a full regression usually takes about an hour. One pilot customer went from 2 days to 1 hour. Your plan sets how many devices and browsers run at once.
Can we start from the regression pack we already have?
Yes. Upload the test cases and specs you already have, connect your repository, and mini explores the app. It writes its tests from all of it, and your team can edit any test in plain words.
We already have automation. Do we throw it away?
No. Keep your unit tests and your scripts. mini runs next to them and takes the end-to-end regression your team still runs by hand, plus the flows your scripts can’t reach.
Can we choose the accounts, codes and devices?
Yes. mini signs in with your test accounts, reads email codes and magic links from its own inboxes, and uses a fixed code you set for SMS. Tests run on iOS simulators and Android emulators, and physical phones are switched on per workspace, on request. mini can also seed test data through your app’s own test endpoints, and reset it with a webhook you set per app.
Where do findings show up?
In the Minitap app, in a Slack thread per issue with Not a bug and Resolved buttons, as a comment on GitHub pull requests to your main or release branches, and as a check on every run. Jira and Linear tickets from an issue are available on request.
What can’t mini test?
Face ID and fingerprint, NFC, Bluetooth, real SMS and calls, real payments, tablets, watches and TV. Payments run in the store’s sandbox or with a test card.
Bring the pack you’re tired of running.
Bring a build and your three hardest flows. A pilot comes first, and you keep the suite.