Frameworks, languages, runners and CI all look like prerequisites. Meanwhile the regression pass you repeat every release keeps taking your week.
For QA engineers
You know where the product breaks. Automation should not start with a framework.
You have tested by hand for years and you know the journeys, the edge cases and the odd data that causes trouble. Now the team wants automation and AI-written tests, and the first step seems to be code you have not learned. This is how CertiQA lets you start from what you know.
- 1Go to/checkout (cart: 1 item, $100)
- 2Type"SPRING20" into Promo code
- 3ClickApply code
- 4Expecttext "20% discount applied" is visible
- 5Expectorder total shows $80.00
Example steps. A test case in CertiQA is a list like this on the canvas.
What this looks like day to day
Four worries we hear from manual testers.
None of them mean you are behind. They are what happens when a good tester is handed tooling built for developers.
You are the person who knows what should be true. It is a poor use of that knowledge to spend months on syntax before you can check anything.
A renamed button or a restructured page fails a test that was fine yesterday, and the morning goes on finding which selector changed.
Assistants can produce a lot of test code quickly. Someone has to judge whether it checks the right thing, and that is hard to do from hundreds of lines of script.
How CertiQA helps
Each worry has a specific answer.
Start by recording a flow you already test by hand
Install the CertiQA browser extension for Chrome or Edge, choose Record Steps and run through the flow in your own browser. Your clicks, typing and navigation appear in the test case as steps. The extension also has a picker, so you can click an element on your live page and send its locators to CertiQA.
A recording captures what you did, not what should be true. You still add the expectations, which is the part you are best placed to write.
You see
- Steps added to your case as you go.
- A canvas where you reorder, edit and add expectations.
- Run history that shows which step failed and why.
Learn Playwright from your own tests, when you want to
Open any web UI test and choose Export to Playwright TypeScript. You get standard @playwright/test code for the steps you built, with options for step comments, network mocks and environment variables. Read how locators, waits and assertions are written for your own application, then edit it if you like.
Reusable sequences such as login become CertiBlocks, so you maintain them once instead of in every case.
How reusable blocks workNot locked in
The exported file is ordinary Playwright code that runs locally or in any CI pipeline. Reading it is optional.
Spend less time on broken selectors
A web step can hold several ways of finding its element, and the step shows how many fallback locator strategies it has. When the primary one stops matching, the alternatives are tried. The Test Health and Flakiness Center lists flaky cases per project, with suggested locator repairs you can look at, and can quarantine tests that fail intermittently.
Healing reduces noise, it does not remove the need to look. Treat a healed step as something to confirm.
How self-healing worksIn the Health Center
- Average flake rate for the project.
- How many tests are quarantined.
- A run action to evaluate quarantine.
Be the reviewer of AI-written tests, not a bystander
Cursor, Claude Code and other MCP-capable assistants can create projects, suites and cases in CertiQA. They arrive as steps on the canvas, so you read what the test does and check it against what you know of the product. You edit it, approve it and only then does it join a suite. The Copilot in the app works the same way: it proposes steps, marks which ones it could not ground in your page, and you approve or reject each one.
Author tests with your AI assistantWhat changes for you
Reviewing intent in plain steps is a skill you already have. You do not need to read generated code to do it.
A path to follow
From manual testing to agent testing, one step at a time.
Automate a flow you test by hand
Pick a regression pass you repeat every release and record it. Add the expectations on the canvas and run it.
Read the code, then reuse it
Export to Playwright and compare it with your steps. Turn repeated sequences into shared blocks.
Test the AI features
Add conversations with mocked tools, an LLM judge and red-team scenarios for any agent your company ships.
The career guide and automation foundations lessons cover each stage in more detail: QA career guide, test automation foundations, example scenarios.
FAQ
Questions manual testers ask about automation
Can a manual QA engineer move into test automation without learning to code first?
You can start without it. In CertiQA a test is a list of plain steps on a visual canvas, so you can build working web UI, API and agent tests from what you already know about the product. Every web UI test also exports to Playwright TypeScript, so you can read the generated code next to your own steps and learn programming at your own pace.
Read the QA career guideDo I need to set up Node, Playwright or a framework?
No. Tests are built and run inside CertiQA. You only need a local Playwright setup if you choose to export a test and run the code yourself.
How do I turn a manual regression pass into an automated test?
Install the CertiQA browser extension for Chrome or Edge, choose Record Steps and walk through the flow in your own browser. Clicks, typing and navigation are added to your test case as steps. You then tidy the steps on the canvas and add the expectations, because a recording captures what you did, not what should be true afterwards.
What happens when the UI changes and a selector breaks?
Each web step can hold several ways of finding its element, shown on the step as a count of fallback locator strategies. When the primary one stops matching, the alternatives are tried, so many cosmetic changes do not fail the run. The Test Health and Flakiness Center also lists flaky cases and suggested locator repairs for you to look at. Healing is a convenience, not a substitute for looking: a healed step is still worth a check.
How self-healing worksCan I export my tests to Playwright?
Yes. A web UI test exports to standard @playwright/test TypeScript. You can choose whether to include step comments, network mocks and environment variables, then run the code locally, commit it or use it in any CI pipeline.
My team uses an AI assistant to write tests. What is my role?
To decide whether the tests are right. Assistants such as Cursor and Claude Code can create cases in your CertiQA project over MCP. They arrive as plain steps you can read, correct and approve before they join a suite, which is where your knowledge of the product matters most.
Author tests with your AI assistantHow do I start testing AI agents?
Describe a conversation as steps: what the user says, which tool responses are mocked and what the agent must or must not do. CertiQA runs it, checks the tool calls and scores the reply with an LLM judge against a rubric you write. You can add failure injection and OWASP LLM Top 10 red-team scenarios later.
AI agent testingIs there a free plan?
CertiQA is in early access: request a spot and we will email you when you can start. Then there is a 14-day trial with no credit card, which includes the Team Pro Plus features and 500 credits. After the trial, access pauses until the account subscribes. See the pricing page for plans.
See pricingPick one flow you repeat every release.
Record it, add the expectations, run it, then export it and read the Playwright code. If you lead a team, see how QA leaders reskill without a hiring cycle.