How long does an exploration take?

The length depends on the size of the app. The exploration agent works across dozens of parallel browsers in the TestMu AI cloud and can run for hours on a large app.

Can I watch an exploration while it runs?

Yes. Live view lets you open any of the parallel browser sessions while the exploration runs.

Does a re-explore start from scratch?

No. A project holds one exploration, and every re-explore adds to the same map. Each run starts from the map of the last one.

Will exploration change data in my app?

Only if you allow it. Read-only mode looks, clicks and validates and never submits a form that changes data. Allow changes fills in and submits forms, so it can reach more screens.

What kinds of findings does Exploration flag?

Exploration flags UX bugs, safety and security issues, and accessibility concerns without being told what to look for.

How do the generated tests get into CI?

A person reviews the tests Kane CLI writes from the map and keeps the ones that match intended behaviour. The approved tests go into CI and run on every future pull request.

Does Exploration work on mobile apps?

Exploration works on web apps today. The rest of Kane CLI tests web apps in real browsers and mobile apps on iOS simulators and Android emulators.

NEW IN KANE CLI · EXPLORATION

Learn your whole app before Kane CLI tests it

Give Kane CLI Exploration a URL. An exploration agent maps every page, control and flow, flags problems, and Kane CLI writes tests from that map.

Try Kane CLI Exploration Read the docs

Automation ExerciseExploring · run #6automationexercise.com · 1 screens · 0 flows · 6 findings

Latest · run #6

Home and help1 screens

Brand products0 screens

Shop and cart0 screens

Account0 screens

Business areas

The problem

Most of the app was never written down, and most of it is never looked at

The spec is missing or out of date

Some apps never had a spec. Others have one that nobody keeps current.

Only the known path gets checked

Tests follow login or checkout, while bugs and screen states hide everywhere else.

Nobody has the full picture

Mapping every page, control and flow is too slow for one person or one browser.

How it works

From one URL to tests in CI, in seven steps

EXPLOREGENERATE AND RUNREVIEW AND SHIP

STEP 1 / 7 · EXPLOREyour terminal

$ kane-cli explore run --url https://automationexercise.com
› Queued on HyperExecute
› Waiting in line for a machine
› Clean profiles, nothing carried over

You point Kane CLI Exploration at the URL of your web app.

Live view

Open any browser while the exploration runs

Dozens of parallel browsers in the TestMu AI cloud explore at once. Pick one to see its screen and what it just did.

automationexercise.com/contact_usLIVE · browser 05

WHERE KANEAI IS EXPLORING›Typing Support User›Typing a test email address›Clicking contact form submit button›Clicking Cart navigation link

What it records · What it flags

Every screen written down, every problem pinned to where it showed up

The Contact Us screen with its recorded controls

29 controls on this screenbanner 9 · form 8 · body 12

Findings across the app65 from runs #5 and #6

Accessibility 55Changed 5Safety 3Coverage 2

Exploration · Assurance loop

Exploration becomes your context

The exploration map feeds the assurance loop, which runs again on every change your coding agent ships.

ExplorationGive it a URL. It explores the whole app on its own, maps every page, control and flow, and flags what is broken.

Start withyour-app.com

Exploration agentParallel browsers, runs for hours

/productsbrowser 01

/view_cartbrowser 02

/contact_usbrowser 03

FindingsUX bugsAccessibilitySecurityApp map

becomes your context

Assurance loopIndependent of the agent that wrote the code. Web and mobile.

Agent drives the change

ContextExploration map, specs and user stories

2 · Kane CLI testsUse cases and acceptance criteria, on web and mobile

3 · Evidence is proofAn Evidence Pack for every run

every
change

QA edge caseSaved as a test. Runs on every future PR.

1 · Agent drivesYour coding agent ships the change

4 · Human signs offSuite stays current as the app changes

Who it's for

For apps nobody fully documented

Legacy apps

Legacy apps whose specs are lost or out of date.

Agent-built apps

Apps a coding agent built faster than anyone could document them.

Inherited products

Products a new team has just inherited.

Hidden coverage gaps

Coverage gaps that your current suites can't see.

Your stack stays

Kane CLI works alongside Playwright, Selenium and Cypress

Your suites keep covering the flows you knew about. Tests from Exploration cover the rest, in the same CI run.

CI · pull request checksall checks passed

✓Playwright suiteknown flows

✓Selenium suiteknown flows

✓Cypress suiteknown flows

✓Kane CLI · tests from Explorationundocumented parts

Questions and current limits

Web apps todayExploration works on web apps today. The rest of Kane CLI tests both web and mobile apps.

Terminal firstA web view for browsing finished explorations is still being built. Explorations start from the terminal today.

A person reviewsA map built from how the app behaves today can include today's bugs, so a person reviews the generated tests.

KaneAI MaxExploration comes with the KaneAI Max plan, which includes three exploration sessions per licence.

Kane CLI Exploration needs the URL of your web app. You start an exploration from the terminal with kane-cli explore run --url and your app’s URL.

Start with the URL of your web app

Install Kane CLI, then point Kane CLI Exploration at your app.

$npm install -g @testmuai/kane-cli

Try Kane CLI Exploration Read the docs