Run Web Automation Tests at Scale with Agentic AI

Generate tests with AI agents or automate test suites with Selenium, Playwright and other frameworks. Run tests in parallel on TestMu AI scalable cloud grid.

Selenium TutorialAutomationTutorial

Selenium Record and Playback: How to Record and Replay Tests

Record a browser test in Selenium IDE, add the checks the recorder skips, and replay the .side file on Chrome, Edge, Firefox, and Safari with selenium-side-runner.

Last Updated on:

A Selenium IDE recording made in March 2025 passed on Chrome, Edge, Firefox, and Safari when I replayed it unchanged on the TestMu AI cloud grid in October 2026. One edited number in its last locator was enough to fail it.

Selenium IDE, the recorder behind record and playback in Selenium, no longer installs from the Chrome Web Store: its listing redirects to an empty page. Record in the Firefox add-on or the desktop app instead, add the checks the recorder skips, and replay the .side file on other browsers with selenium-side-runner.

Overview

Record and playback in Selenium runs through Selenium IDE. It records each click and keystroke as a command, target, and value row, replays the rows in the browser, and saves them as a .side file. As of October 2026, install it as a Firefox add-on or desktop app, because the Chrome Web Store listing is empty.

Where Can You Install Selenium IDE?

  • Firefox add-on: Selenium IDE 3.17.4 is still listed on Firefox Add-ons, where it installs in Firefox and records in the 3.x interface.
  • Desktop app: Selenium IDE 4.0.1-beta.14 is a standalone app for Windows, macOS, and Linux, released in July 2024 with a version name that still carries the beta label.
  • Chrome Web Store: Not available; the Selenium IDE listing that selenium.dev links to redirects to an empty page, so Chrome users record in Firefox or the desktop app and replay the file in Chrome.

How Do You Replay a Recording on Other Browsers?

  • selenium-side-runner: The npm command-line runner, version 4.0.13, replays a .side file against a local browser driver or a remote Selenium Grid, with browser and platform set through its -c capability flag.
  • Cloud replay: On the TestMu AI grid, a Selenium IDE recording made in March 2025 passed unchanged on Chrome 154, Edge 154, Firefox 156, and Safari 18 in October 2026.

What Does a Recording Not Check?

  • Assertions: Selenium IDE records actions but adds no checks, so a replay that passes proves only that each locator matched; add assert or waitFor rows for the outcome you expect.
  • Alternate locators: Each recorded row stores alternate locators, but selenium-side-runner 4.0.13 tried only the primary target when a recorded id stopped matching the page.

What Is Record and Playback in Selenium?

Record and playback testing is test automation in which a tool captures the actions a person performs in an application and replays them as an automated test. In Selenium, the recorder is Selenium IDE: it saves each action as a command (such as click or type), a target (the element's locator), and a value (the text typed or the key pressed).

Selenium IDE is the only part of the Selenium project that records. Selenium WebDriver is a library you call from code, and Selenium Grid runs WebDriver sessions on remote machines; neither records, but a Grid can replay an IDE recording through selenium-side-runner. The Selenium IDE tutorial covers the IDE's other features, and the guide on how to use Selenium covers WebDriver from setup to locators. If you write Playwright tests instead, the Playwright recorder generates code rather than command rows.

A recording is saved as a .side file, a JSON project that lists each test's commands. The test used throughout this guide searches the eCommerce Playground for imac and opens a product page; Tahera Alam recorded it for the guide's first version in March 2025.

These are its rows, copied from the original .side file:

RowCommandTargetValue
1openhttps://ecommerce-playground.lambdatest.io/(none)
2setWindowSize1440x778(none)
3clickname=search(none)
4typename=searchimac
5sendKeysname=search${KEY_ENTER}
6clickcss=#mz-product-grid-image-41-212469 > .carousel-inner > .active > .lazy-load(none)

Rows 3 to 6 also store the locators the recorder saw, in a targets list. Row 6 lists four: the CSS path above and three XPath expressions, one of them xpath=//img[@alt='iMac'].

How to Record and Play Back a Test in Selenium IDE

The steps below produce the recording above. The screenshots show Selenium IDE 3.17.2; the Firefox add-on ships 3.17.4 from the same 3.x line.

Install Selenium IDE

Selenium IDE no longer installs in Chrome, so use the Firefox add-on or the desktop app:

  • Firefox add-on - Selenium IDE on Firefox Add-ons is version 3.17.4, the 3.x interface in this guide's screenshots. Click Add to Firefox, then open the IDE from its toolbar icon.
  • Desktop app - The Selenium IDE releases page on GitHub offers 4.0.1-beta.14, published in July 2024, as a Windows installer, a macOS disk image, and a Linux AppImage. Its version name still carries the beta label.
  • Chrome extension - The Chrome Download button on selenium.dev leads to a Chrome Web Store listing that redirects to an empty page. In Chrome, use the built-in Chrome DevTools Recorder instead; TestMu AI's guide to record and replay tests using LT Debug shows how to replay those recordings on a cloud grid.

With the add-on or the app installed:

  • Open Selenium IDE and click Create a new project in the welcome dialog.Selenium IDE 3.17.2 welcome dialog with the Create a new project option
  • Name the project, here record-playback-demo, and click OK.Selenium IDE dialog naming a new project record-playback-demo

Create a Test

A new project opens with one empty test named Untitled. Click the + button (Add new test) above the test list, name the test first-test, and click ADD.

Empty Selenium IDE project with the Add new test button highlighted above the test list

Record the Test

  • Enter https://ecommerce-playground.lambdatest.io/ in the Playback base URL field.Selenium IDE playback base URL field set to the eCommerce Playground
  • Click REC at the top right. Selenium IDE opens the base URL in a new browser window and shows a Selenium IDE is recording banner.eCommerce Playground home page open in a browser window with the Selenium IDE is recording banner
  • Click the search box, type imac, and press Enter. The results page loads while the IDE records each step.eCommerce Playground search results for imac while Selenium IDE records
  • Click an iMac product to open its page.iMac product page on the eCommerce Playground while Selenium IDE records
  • Return to the Selenium IDE window and click the record button again to stop recording.

The rows in the table above are the result. Selenium IDE added the open and setWindowSize rows itself when recording started.

Play Back the Test

Click Run current test in the toolbar. Selenium IDE replays each row in a browser window and marks it green when it passes.

Selenium IDE toolbar with the Run current test button highlighted above the six recorded rows

The Run all tests button to its left replays every test in the project in order.

Selenium IDE toolbar with the Run all tests button highlighted

The Log panel lists each executed command with OK and ends with 'first-test' completed successfully. A failing row turns red, and the log shows its error, such as an element that could not be found.

Selenium IDE log panel showing all six commands OK and first-test completed successfully

Add Assertions the Recorder Skips

The recording has no row that checks a result, so its replays pass whenever each locator matches an element. Add two rows below row 6: select the empty row under the last command and fill in the Command, Target, and Value fields under the table.

RowCommandTargetValue
7waitForElementPresentxpath=//h1[text()='iMac']15000
8assertTitleiMac(none)

Row 7 waits up to 15 seconds for the product heading, and row 8 fails the test if the page title is anything other than iMac. With both rows added, the replay passed on all four browsers in the cloud run below.

The playground lists several products named iMac, and the two I compared share the same page title and product code, so this check confirms that an iMac product page opened; it cannot tell those products apart.

Export the Test to WebDriver Code

Selenium IDE exports a test as WebDriver code you can keep in a repository. Open the test's menu (the three dots next to its name) or right-click it, and choose Export.

Selenium IDE test list with the three-dot menu next to first-test highlighted

Pick C# NUnit, C# xUnit, Java JUnit, JavaScript Mocha, Python pytest, or Ruby RSpec, and click Export. The 3.17 dialog can also add origin tracing comments, add each step's description as a comment, and prepare the file for a Selenium Grid.

Selenium IDE Select language export dialog listing C#, Java, JavaScript, Python, and Ruby options

The exported file is the start of a coded test. The .side file stays the one selenium-side-runner replays.

How to Replay Selenium IDE Tests on a Cloud Grid

selenium-side-runner replays a .side file from the terminal against any WebDriver endpoint, so one recording runs on browsers you have not installed. Version 4.0.13, the newest on npm in October 2026, needs Node.js 16 or later. I used it to run Selenium IDE tests on the cloud grid at TestMu AI.

Install the runner:

npm install -g selenium-side-runner

Save the project in Selenium IDE with the disk icon at the top right to get record-playback-demo.side.

Selenium IDE with the save project icon highlighted at the top right

Copy your TestMu AI username and access key, then run these lines in bash (macOS, Linux, or Git Bash on Windows) from the folder that holds the file:

export LT_USERNAME="your-username" LT_ACCESS_KEY="your-access-key"
HUB="https://$LT_USERNAME:$LT_ACCESS_KEY@hub.lambdatest.com/wd/hub"
CAPS="browserName=chrome browserVersion=latest"
CAPS="$CAPS LT:Options.platformName='Windows 11'"
CAPS="$CAPS LT:Options.build='Selenium IDE replay'"
selenium-side-runner record-playback-demo.side --server "$HUB" -c "$CAPS"

The -c flag takes space-separated key=value pairs, and a dotted key such as LT:Options.platformName becomes a nested capability. For Edge or Firefox, change chrome in the third line to MicrosoftEdge or firefox. For Safari, use browserName=Safari and set platformName to 'macOS Sequoia'.

Capability code from a language binding does not work in -c. When I passed FirefoxOptions() lines asking for Firefox on macOS Sequoia, the runner did not apply them, and the session started as Chrome 154 on Windows 11 in an untitled build. To look up browser, version, and platform names, use the Automation Capabilities Generator and copy each value into a key=value pair.

These were the results on the TestMu AI grid on October 12, 2026: the unchanged recording in build 108177063, and the commands above, with the two assertion rows added, in build 108183375.

BrowserPlatformUnchanged recordingWith rows 7 and 8
Chrome 154Windows 11PassedPassed
Edge 154Windows 11PassedPassed
Firefox 156Windows 11PassedPassed
Safari 18.0macOS SequoiaPassedPassed

Trimmed to its result lines, the Chrome run printed:

info: Running test first-test
info: Finished test first-test Success
    √ Running test first-test (19582 ms)
Tests:       1 passed, 1 total

Each run appears in the TestMu AI Web Automation dashboard with a video and a command log. The view below is from the first version of this guide, when the same recording ran on Chrome 133 and Windows 11.

TestMu AI Web Automation dashboard showing a completed Selenium IDE replay with its command list and video
Note

Note: Replay your Selenium IDE recordings on 3,000+ browser and OS combinations with TestMu AI. Try TestMu AI free

Limitations of Record and Playback Testing

Recorded tests break when a captured locator stops matching the page, and they pass without checking anything until you add assertions. This guide's recording shows both.

  • Locators that encode data - Row 6's id, mz-product-grid-image-41-212469, combines the product's id (41) with a number every card in the results grid shares (212469). Two of the original screenshots, the save and export ones, show another take of the same step recorded as product 54: the search returns several products named iMac, and the recorder stores the id of whichever card you click.
  • Passing without checks - The unchanged recording passed on four browsers on the strength of its locators alone; no row checked a result.
  • Alternates the runner skips - A row's recorded alternates did not help when its primary locator stopped matching, as the test below shows.
  • Upkeep when the flow changes - A changed flow means editing or re-recording every test that recorded it, unless the shared steps live in one test that others call with the run command.

selenium.dev says Selenium IDE "records multiple locators for each element it interacts with. If one locator fails during playback, the others will be tried until one is successful." To test that claim in the runner, I mimicked a redesign by changing 212469 to 999999 in row 6's target and left the row's recorded alternates in place.

On the TestMu AI grid, the session's command log shows 48 lookups of the edited CSS selector and none of the XPath alternates before selenium-side-runner 4.0.13 failed the step. The edited copy ran as a test named redesign-primary-broken:

info: Running test redesign-primary-broken
info: Finished test redesign-primary-broken Failure
    × Running test redesign-primary-broken (31995 ms)
    Operation timed out during click:css=#mz-product-grid-image-41-999999 > .carousel-inner > .active > .lazy-load:
Tests:       1 failed, 1 total

With the recorded xpath=//img[@alt='iMac'] alternate promoted to row 6's primary target, the test passed again on Chrome 154. That locator clicks whichever iMac card comes first, which is fine when any iMac page will do. The no-code browser automation guide traces the skipped alternates to the runner's source code.

Austin Siewert

Austin Siewert

Co-Founder, Steadfast Systems

Discovered @TestMu AI yesterday. Best browser testing tool I've found for my use case. Great pricing model for the limited testing I do 👏

3M+ Devs and QAs rely on TestMu AI

Deliver immersive digital experiences with Next-Generation Mobile Apps and Cross Browser Testing Cloud

Best Practices for Selenium Record and Playback

Most of these are edits in the Command, Target, and Value fields before you save the .side file.

  • Check every target before you save - Prefer hand-written ids, name attributes, or data-* attributes over ids generated from data. The name=search target in rows 3 to 5 still matched in October 2026, 19 months after recording, while one changed number in row 6's data-derived id broke the test. This Selenium locators tutorial compares the options.
  • End each test with an assertion - Add an assert or verify row for the result you would check by eye, such as the page title, a heading, or a confirmation message.
  • Wait with waitFor rows, not pause - waitForElementPresent and waitForElementVisible take a timeout in milliseconds and move on as soon as the element appears, while pause always waits the full time.
  • Record shared steps once - Keep a login or search flow in its own test and call it from other tests with the run command, so a changed form needs one edit.
  • Use control flow sparingly - Rows such as if, while, and times handle pages that vary, but a test that needs heavy branching is easier to maintain as exported WebDriver code.
  • Replay on every browser your users have - A pass in the IDE covers only the browser it ran in, and selenium-side-runner with a different browserName in -c covers the rest.

How Do AI Agents Change Record and Playback Testing?

AI testing agents tie each step to what you meant rather than to one recorded selector: you show or describe the flow once, and the agent finds the element on each run.

  • Recording into natural language - KaneAI, TestMu AI's testing agent, has a Manual Interaction mode that captures each click, typed value, scroll, or navigation and turns it into a natural language test step, as the guide to Manual Interaction in KaneAI describes. Recording can pause and resume, and KaneAI's smart element detection identifies the intended element by intent and context instead of a fixed CSS or XPath selector.
  • Export to frameworks - KaneAI exports tests to Selenium, Playwright, Cypress, or Appium code, much as Selenium IDE exports WebDriver code.
  • Limits to know - KaneAI's documentation lists hover actions that open tooltips or menus as not recordable on desktop web, and only the locator name of a recorded step can be edited afterwards.
  • Coding agents - Claude Code or GitHub Copilot can write a Selenium WebDriver test from a plain English description of the flow. The guide on how to use the Selenium skill with Claude Code shows how a SKILL.md file makes the agent use stable locators, explicit waits, and Page Objects in the code it writes.
  • The review step stays - An agent-written or agent-driven test still needs a person to confirm that its assertions match the intended behavior, and a Selenium IDE recording is still the quicker choice for a one-off check like the iMac search in this guide.

Skip the setup and install the Selenium Skill for Claude Code, Copilot & Cursor with one command.

Selenium

Conclusion

Start with one flow you already test by hand: record it in the Firefox add-on, add an assert row for the result you check by eye, and replay the .side file with selenium-side-runner. To run that replay across browsers without a local lab, point the runner at the TestMu AI test automation cloud, which covers 3,000+ browser and OS combinations; the guide to running Selenium IDE tests on TestMu AI explains parallel runs.

Testing web or mobile apps without writing code? KaneAI by TestMu AI turns plain-English instructions into tests.

To run your first test, start free, or
for a full walkthrough of the product.

Author

...

Navin Chandra

Blogs: 5

  • Linkedin

Navin Chandra is a Member of Technical Staff at TestMu AI, building the open-source automation that powers its Selenium and Appium cloud grid. A committer to both Selenium and Appium, he implemented WebDriver BiDi support in Selenium for real-time browser events and bidirectional control and is developing Apple's iOS RemoteXPC protocol in Appium to enable low-level wireless communication with iOS system services. He contributes to Selenium across multiple language bindings as a member of the Selenium GitHub organization. He has served as a Google Summer of Code mentee and mentor at openSUSE and an LFX mentee at CNCF's KubeArmor, and is a SUSE Certified Deployment Specialist. Navin holds a B.Tech in Computer Science.

Reviewer

...

Sri Harsha

Reviewer

  • Linkedin

Sri Harsha is Engineering Manager of the Open Source Program Office at TestMu AI, where he leads open-source engineering behind the Selenium and Appium automation grid and builds agentic AI systems for quality engineering. He is a member of the Selenium Technical Leadership Committee and a committer to WebdriverIO and Appium, and was recognized with the LambdaTest Delta Award 2023 for Best Contributor in open-source testing. He brings over 10 years of experience in software testing and automation, with earlier roles at EPAM Systems and ZenQ. Sri Harsha holds a B.Tech in Computer Science from Jawaharlal Nehru Technological University.

Add to Google preferred sources

Summarise with AI

Copied to Clipboard!
...

3000+ Browsers. One Platform.

See exactly how your site performs everywhere.

Try it free
...

Write Tests in Plain English with KaneAI

Create, debug, and evolve tests using natural language.

Try for free

Selenium Record and Playback FAQs

Did you find this page helpful?

More Related Blogs

TestMu AI forEnterprise

Get access to solutions built on Enterprise
grade security, privacy, & compliance

  • Advanced access controls
  • Advanced data retention rules
  • Advanced Local Testing
  • Premium Support options
  • Early access to beta features
  • Private Slack Channel
  • Unlimited Manual Accessibility DevTools Tests