1. Docs
  2. Use cases
  3. Use cases
  4. How do I automate a web test without writing code?

How do I automate a web test without writing code?

In SmartestQA you record a web test instead of writing it: click Record in Automate, do the flow in your app in a new tab, and the browser extension adds every click, typed value and selection as a step. You then add a check with Assert from the step toolbar and try the test with Dry Run.

11 min read Updated CustomersAdmins

In short

On this page you can

  • Record clicks, typed values and selections in Chrome or Edge as test steps.
  • Add checks, API calls, database queries and mail steps from the step toolbar, without code.
  • Compare a value with what you expect, using one of ten assertion types.
  • Try the test with a dry run and watch every step live, with a screenshot.
  • Reorder, switch off, remove or rebind steps in the editor.

How to use it

  1. Install the SmartestQA browser extension and add a start page URL in Configuration › Project Panel › Start Page URLs.
  2. In Automate click New Test Case, pick the start page URL and the environment.
  3. Click Record, do the flow in the new tab, close the tab and come back.
  4. Add an Assert step at the end and click Save.
  5. Click Dry Run, choose Run on Server or Run on your Agent and read the live log.

The short answer#

You do not write a script. In a SmartestQA web project you open a test case in Automate, click Record and use your application in a new browser tab. The SmartestQA extension for Chrome and Edge adds every click, typed value and selection as a step, and creates one action per control with its locators and a screenshot. When you close the tab, the steps are in the test case.

What you cannot click, you add from the step toolbar: a check with Assert, an API call, a Database query, a Mail step. Dry Run then runs the saved test once and shows each step live with a screenshot. Code steps (Python, JS) exist for what no step type covers.

The problem this solves#

In many teams the people who know the application best are testers and business users who do not write code. They check the important flows by hand before each release, and they repeat the same clicks every time. The checks depend on who has time, and nothing records what was checked.

Automating a flow usually means a script, so it waits for a developer or a test engineer. When a page changes, the script waits for that person again. What is missing is a way for the person who knows the flow to turn it into a test that runs again, and to see it run.

An example#

This is an illustration, not a customer story. A team runs an HR portal, and a tester who has never written a script wants to automate the flow “submit a leave request”.

  1. The tester installs the SmartestQA extension in Chrome. A start page URL with a TEST Environment URL for the portal is added under Start Page URLs.
  2. In Automate the tester clicks New Test Case, picks the start page URL and the environment, and clicks Record. The portal opens in a new tab. The tester logs in with a test user, fills in the leave request, submits it and closes the tab. SmartestQA shows “9 new step(s) added!”.
  3. The flow is recorded, but nothing is checked yet. The tester adds a step with an action of type Get that stores the confirmation message, then clicks Assert, chooses the type Contains and sets the expected value submitted.
  4. The tester names the test case, clicks Save, then Dry Run and Run on Server. The Test Run panel opens with the live log and a screenshot for every step. It ends with “Test completed successfully!”.
  5. A dry run is not stored in a report, so the tester adds the test case to a pipeline in Schedule and clicks Save & Execute. The result is now in Report.

The person who knows the flow built the test, and no script was written.

By hand, with hand-written test scripts, and with SmartestQA#

QuestionChecked by handHand-written test scriptsSmartestQA
Who can build the test?Anyone who knows the applicationSomeone who can programAnyone who can do the flow in Chrome or Edge with the extension installed
How does a click become a step?It does not; a person clicks again each timeYou write a locator and a command for each controlThe recorder adds every click, typed value and selection as a step
How is the result checked?By eye, which works well for a new or rarely used flowWith an assertion in codeWith an Assert step that compares a saved value with the expected one
What do you see while it runs?The application itselfWhat the script logsA live log, with a screenshot for every step of a dry run
Where does it run?On the tester’s machineOn a machine or CI runner you set upRun on Server, headless with no agent, or Run on your Agent on your computer
What if no ready-made step fits?The tester does it by handYou write more code, with full controlPython and JS code steps exist for that case

Steps#

  1. Install the browser extension

    Open SmartestQA in the Chrome Web Store and click Add to Chrome. The same listing works in Edge. Then press F5 on every SmartestQA tab that was already open: a newly installed extension cannot talk to tabs that were open before it. See Install the browser extension.

  2. Add a start page URL

    Go to ConfigurationProject PanelStart Page URLs and click + Add. Give it a Name and at least one of DEV Environment URL, TEST Environment URL, PRP Environment URL. Addresses must start with http:// or https://. See Add a start page URL.

  3. Open a test case and pick where it starts

    Go to Automate and click New Test Case. In Test Case Detail pick the start page URL and the environment. Without a start page, recording does not start: “Please select a Start Page URL.” See Environment and start page.

  4. Record the flow

    Click Record at the left of the step toolbar. Your app opens in a new tab. Log in with a test user and work through the flow the way a user would. Wait for each page to load before you click. Then close the recording tab and return to SmartestQA.

    “n new step(s) added!” The new steps are appended at the end of the list. Each one points at an action in Discover.

  5. Add a check

    A test proves something only when it checks a result. Click Assert in the toolbar; the dialog Add Assertion/Verify opens. Fill Page Name, Assertion Type and Test Step Name, click Add, then set actual and expected in the right panel. An assertion compares values and does not read the page, so save the value first with an action of type Get. See Add an Assertion step.

  6. Name and save

    Give the test case a name and click Save. You do not have to write a description: the AI writes one from the steps. There is no autosave. See Save.

    “name updated successfully”.

  7. Dry run it

    Click Dry Run, just left of Save, and choose the run target: Run on Server (headless, no agent needed) or Run on your Agent (on your computer; you can watch the browser). See Web projects.

    “Test execution started successfully” and the Test Run panel opens with the live log. It ends with “Test completed successfully!” or “Test failed”.

  8. Keep the result

    “Dry Run results are not saved to reports.” For a result you can keep and share, go to Schedule, create a pipeline, add the test case and click Save & Execute. Then open Report. See Run it and read the report.

Steps you can add without code#

These are buttons of the step toolbar in a WEB project that need no code. The full list is in The step toolbar.

ButtonAddsGood to know
RecordA recorded flow: clicks, typed values, selectionsWeb recording saves at once, as you record.
ActionA step bound to an existing actionThe Action List dialog stays open, so each click adds one step.
TemplateA template test case as one stepChange the template and every test case that uses it changes too.
AssertA comparison of an actual value with an expected oneA failed assertion stops the test case.
AI stepA step written in plain language that the AI performs, answers or checksCalls the AI on every run: slower, not deterministic, and needs AI turned on.
APIAn HTTP request, with values extracted from the responseAny HTTP status passes the step; extract @status and assert it if the code matters.
NetworkA check of the browser’s own network callsWEB projects only.
More…DatabaseA SELECT query whose columns become variablesThe query runs from where the test runs.
More…MailRead mails, extract an OTP or link, send or wait for a mailReads only unread mails that arrived shortly before or after the step starts.

The ten assertion types#

The Assertion Type list shows the types set up for your installation. By default these ten. Text is compared case-sensitive and is not trimmed; see How values are compared.

TypePasses whenExample (actual → expected)
EqualsThe values are exactly equal.Paid → Paid
Not EqualsThe values are not equal.Draft → Paid
ContainsThe value includes the expected text.Order 1042 created → created
Not ContainsThe value does not include it.Saved → Error
Starts WithThe value starts with the expected text.INV-2026-17 → INV-
Ends WithThe value ends with the expected text.report.pdf → .pdf
Is EmptyThe value is empty or blank.No expected value
Is Not EmptyThe value is not empty.No expected value
Greater ThanThe value is a number greater than expected.1234.5 → 1000
Less ThanThe value is a number less than expected.850 → 2000

What you get#

  • A test case built from what you did. One step per click, typed value and selection, each linked to an action with locators and a screenshot.
  • Checks without code. Ten assertion types by default, plus a Visibility Check action that verifies an element or a text is visible on the page.
  • A live log. One row per active step with a status badge, messages, duration and a screenshot; dry runs take a screenshot for every step.
  • Two places to try it. Run on Server runs headless with no agent; Run on your Agent opens a visible browser on your computer.
  • An editor for the steps. Drag to reorder, Deactivate a step, Remove it, or bind it to another control with Replace Action.
  • Reuse. A recorded action can be added to any other test case with Action, and a template test case can be added as one step with Template.
  • Locator repair. When all locators of a control fail in a run, auto-heal looks for the control; if it finds it, a review is created in Maintenance.

Questions and answers#

Do I need to know programming to automate a web test in SmartestQA?Customers

No. You record the flow with Record in Automate and add checks, API, database and mail steps from the step toolbar. Python and JS steps are there for what no step type covers, and you can leave them out. See Which step family to use.

Which browsers can I record in?Customers

Chrome and Edge. The SmartestQA extension is installed from the Chrome Web Store with Add to Chrome, and the same listing works in Edge. Without the extension, Record opens the Web Store instead of your app. See Install the browser extension.

Do I have to install anything besides the browser extension?Customers

Not for a web app that is reachable from the internet: Run on Server runs headless on the Smartest Server and needs no agent. You need an agent when you want to watch the browser during a run, or when the app is behind a VPN, a firewall or on an intranet. See Do you need an agent?

How do I check that the page shows the right result, without code?Customers

Read the value with an action of type Get, which stores the text under a key you name, then add an Assert step that compares it with the expected value, for example Contains saved. To verify only that an element or a text is visible, use an action of type Visibility Check. See Common patterns.

Can I record a test that starts with a login?Customers

Yes. Log in in the recording tab and go on with the flow. Use a test user: the values you type are stored in the steps, and recording saves a screenshot of every control you use. Values typed into fields named like password or token are masked in run logs but still stored in the step, so put real secrets in a data source. See Secrets.

Can I change the steps after recording?Customers

Yes. In the editor you drag a step to reorder it, use Deactivate to leave it out of every run, Remove to delete it, and Replace Action to bind it to another control. Typed values are changed in Parameters List. These changes are kept when you click Save. See Change steps.

Are recorded steps saved automatically?Customers

Steps recorded with the web extension are saved at once, while you record. Everything else you change in the editor is saved only when you click Save; there is no autosave and no keyboard shortcut for it. See What is saved without Save.

Does a dry run give me a report?Customers

No. A dry run shows a live log, and its screenshots are deleted after 1 hour. A pipeline run from Schedule is stored in Report with screenshots, video and logs. See Run modes: dry run, pipeline, agent.

What do I do when a recorded step fails?Customers

Open the first red step in the live log; later steps are usually SKIPPED because of it. Read the message, look at the screenshot, change the value, the wait, the order or the locator, save and dry run again. See Diagnose a failed step.

Do I have to map my application before I record a test?Customers

No. Recording in Automate creates the actions its steps use, so the test works. What you lose without Discover is context: the App Map has no purposes or coverage, and AI-generated requirements and tests know only the names of controls. See Why spend time here.

Read more#

Didn’t find your answer?Write to support with your project name and what you see on screen.

info@smartestqa.ai