# Tester: Clear PASS or FAIL Results With Screenshots

> The Tester is the AI website testing agent in the Ai1 platform by MyZone AI that opens a real browser, steps through your site or app at desktop and mobile screen sizes and reports every check as a numbered PASS or FAIL with screenshots, without changing the work it checks.

Canonical page: https://myzone.ai/pages/agents/ai-tester-agent
Part of Ai1, by MyZone AI. Book a Tester walkthrough: https://calendly.com/d/ct6h-tcy-8qf/ai1-demo?a1=Tester&utm_source=myzone.ai&utm_medium=agent-page&utm_content=ai-tester-agent-final
Last updated: 2026-10-05. Reviewed by the MyZone AI team.

Every check comes back PASS or FAIL, with the screenshot to prove it.

## At a glance

- **You give it:** The page address and what you expect to work
- **Screen sizes:** Desktop and a mobile size of 375 by 812 pixels, in a real browser
- **What it checks:** Layout, buttons, forms and links, your automated tests and every fix
- **On request:** Core Web Vitals, page speed and WCAG 2.2 AA accessibility
- **Each result:** A numbered PASS or FAIL, with expected versus actual for every failure
- **What it never does:** Fix product code: failures go back to your developer

## What the Tester does

- **Checks pages in a real browser:** Real browser testing: it sets the screen size, opens the page and checks that it looks right and that buttons, forms and links work, on both mobile and desktop.
- **Runs your automated tests:** It runs your project's test suite and reports every result. Anything that fails comes with the full error output, so your developer knows exactly what to fix.
- **Sweeps for knock-on breaks after a fix:** Once a bug is fixed, it confirms the fix and then checks the features around it to make sure nothing nearby broke.
- **Reports with evidence:** Every verdict lists numbered test cases, shows expected versus actual for each failure and includes screenshots, so there is no doubt about what passed.
- **Checks speed and accessibility:** Before a launch or after a redesign, it can run Core Web Vitals and page speed checks and audit pages against WCAG 2.2 AA accessibility.
- **Writes new test code when you need it:** If a check needs a new automated test, it writes one in your project's test framework, runs it and reports the result. It writes test code, never product code.

## How the Tester differs from a quick look in your own browser

A quick look in your own browser tells you a page seems fine on one screen. The Tester sets the screen size, steps through the page like a visitor at desktop and mobile sizes and records a numbered PASS or FAIL for every check, with screenshots and expected versus actual for each failure. After a fix it sweeps the features nearby, and it never changes the work it checks.

How this differs from the QA Orchestrator (https://myzone.ai/pages/agents/ai-qa-agent): the QA Orchestrator decides how deep to check a change and combines the Tester's results with other specialists' into one verdict, while the Tester runs the browser checks and test suites.

## How it works

1. **Take the brief** (When a change is ready): You share the page address, the behaviour you expect or the test to run. If anything is unclear, it asks before starting.
2. **Set up the session** (Before the first check): It sets the right screen size and, for browser checks, reads the acceptance criteria from your project documents.
3. **Run the checks** (Case by case): It navigates, clicks, fills in forms, runs tests and captures screenshots, recording a result for every case.
4. **Compile the report** (When the checks finish): It closes the browser and builds a numbered PASS or FAIL report with screenshots and error details.
5. **Hand back the results** (Before the work counts as done): You get the full results. When several checks run together, the QA Orchestrator combines them into one verdict, and failures go to your developer.

## When to use it

- **A developer says a change is ready:** It steps through the expected behaviour in a browser and returns a verdict with screenshots before you mark the work done.
- **You want to see how a page looks on mobile:** It switches to a mobile screen size, checks the layout and interactions, and captures a screenshot.
- **A bug was fixed and you want to know nothing else broke:** It confirms the fix, then checks the surrounding features at both mobile and desktop sizes.
- **You want your automated test suite run:** It runs the tests, reports every result and includes the full error output for anything that fails.
- **A launch or redesign is coming up:** It runs speed and accessibility audits and gives you the results to act on before visitors arrive.

## What you get

- A numbered PASS or FAIL list covering every check
- Screenshots at desktop and mobile screen sizes
- Expected versus actual for each failure
- Full error output from automated test runs
- Core Web Vitals, page speed and accessibility results when requested

## Example: booking page release check

Example with a fictional company. Names, people and figures are invented to show the agent's output. Any resemblance to a real company or person is unintended.

What it was asked: Check our booking page on desktop and phone before release, and tell us which issues must be fixed first.

Tester ran six deterministic browser checks on one disposable synthetic local booking page at desktop and phone widths, and captured the two rendered views. The three failures are real observations of deliberately seeded faults in that fixture, not claims about a client site; the front-end confirmation did not send a booking to a server. No production system or personal data was touched, and the browser was closed. Run date: 30 September 2026.

### What it found

- Submitting without a preferred date shows “Request sent” instead of stopping the request.
- The service card says CAD $95, while the booking summary says CAD $80 for the same selection.
- At 375 px wide, page content measures 759 px, leaving 384 px of horizontal overflow.

## Guardrails

- It checks the work and never changes it. Product code fixes go back to your developer, which keeps the check independent.
- If the brief is too vague for a reliable check, it asks one or two questions instead of guessing.
- Mobile checks run in a real browser set to a mobile screen size, not on physical handsets.
- Deciding what to build and what counts as done stays with you and your team.

## Frequently asked questions

### Why test a website in a real browser?

Because that is where visitors meet it. Browser testing opens the page, sets the screen size and checks that it looks right and that buttons, forms and links work. The Tester in Ai1 by MyZone AI does this at desktop and mobile sizes and records a numbered PASS or FAIL for every case, with screenshots. For mobile it uses a real browser set to a screen size of 375 by 812 pixels, which catches the vast majority of layout and interaction problems. It does not run on physical devices.

### What should a pass or fail test report include?

A numbered list of every check with its result, expected versus actual for each failure, screenshots at the sizes tested and any raw error output. That is what the Tester hands back, so your developer can start fixing at once. The Tester does not repair what it finds. Its job ends at the verdict and the evidence, which keeps the check independent.

### Can automated QA testing check page speed and accessibility?

Yes. Before a launch or after a redesign, the Tester can run Core Web Vitals and page speed checks and audit pages against WCAG 2.2 AA accessibility, alongside its browser checks and your automated test suite.

### What if I am not sure what to test, and can the Tester write new tests?

Tell it what you changed or what worries you. If that is too vague for a reliable check, it asks one or two questions first. It never guesses. If you need a new automated test for your project's test framework, it writes it, runs it and reports the result.

## More on this topic

- [Your Website Is Getting Slower Every Day. You Just Don’t Know It Yet.](https://myzone.ai/pages/blog/ai-pagespeed-monitor): Sites get slower with every update. Daily automated PageSpeed monitoring catches regressions before they cost you rankings and conversions.
- [Multi-Agent Development: Why the Future of Software Is Built by AI Teams](https://myzone.ai/pages/blog/multi-agent-development): Discover how multi-agent AI systems are transforming software development. Teams of specialised AI agents build products 10x faster at lower cost.

## About Ai1

Ai1 is the AI operations platform by MyZone AI, where each client runs on its own private server. The Tester is included on every Ai1 level, including Developer Core, with no per-agent charge. It works alongside the other Ai1 agents on your account.

Pricing: https://myzone.ai/pages/services/ai1-pricing. Security: https://myzone.ai/pages/security.