For QA testers and developers
You click. TestMaadu writes the end-to-end test code.
- 1extension
- 6QA tools
- 0AI tokens
Stop hunting for locators and typing the same assertions again. Pick an element, record a bug or capture an API call, and get ready-to-run Playwright, Cypress or Selenium code. Bug reports, JMeter load tests, security checks and AI chatbot tests come from the same browser extension.
Or try it yourself, right here in your browser- Free to try, no card
- Your data stays in your browser
Sound familiar?
“I spend more time finding locators than writing tests.”
Flaky XPaths, IDs that change every build, and the same expect() lines in test after test.
“Can't reproduce. Works on my machine.”
The report missed a step, the console error or the failed API call, so the bug bounces back and forth.
“We'll do API, load and security testing later.”
Each needs its own tool and setup, so “later” never comes before the release.
“Our new AI chatbot is live. How do we even test it?”
The answer changes every time, so normal assertions break. Nobody checks whether it leaks another customer's details, quotes the wrong price or obeys “ignore your instructions”.
You test manually. TestMaadu records the steps and gives you the automation code, in just a few clicks.
TypeScript·JavaScript·Python·Java
Try it yourselfYour QA companionNo AI tokens needed
One extension. Six jobs done for you.
Click a card to see everything it does, and how it helps you as a tester.
Point at any element and get a locator and code that work the first time.
| What it does | How it helps you as a QA |
|---|---|
| Stable locators, ranked by how robust they are | Fewer tests break on the next build, because IDs that change and long XPaths are avoided. |
| Verify on the page before you copy | You see the click or the typing happen, so you never paste a locator that finds nothing. |
| Confidence rating and “Why this locator?” | You can defend the locator in a code review, and newer testers learn what a good locator looks like. |
| Freeze (Alt+Shift+F) for toasts, tooltips and menus that vanish | Elements that disappear before you can inspect them get a locator too. |
| Iframes, shadow DOM, tables, custom dropdowns, date pickers and uploads | The hard parts of real enterprise apps work without hand-written workarounds. |
| Row and card anchors for repeated items | “The View button in order #1042’s row” instead of the third View button, so tests survive new rows. |
| Playwright, Cypress and Selenium in TypeScript, JavaScript, Python and Java | The code fits the suite your team already runs. No conversion. |
| Find all: every key element, then a Page Object, screen test or CSV | A new screen’s Page Object takes minutes, not an afternoon. |
The free trial unlocks all of it. Pricing shows what stays free after the trial.
Record a bug once, and developers can reproduce it on the first try.
| What it does | How it helps you as a QA |
|---|---|
| Steps written as you test, with the test data used | Saves the write-up time on every bug. Nothing is forgotten. |
| Console errors and failed API calls tagged with the step where they happened | Developers see the cause next to the action. This ends most “can’t reproduce” replies. |
| API errors with request, response and Copy as cURL; slow calls | Developers replay the failing call themselves, and backend bugs go to the right team. |
| Screen recording of the tab, with your voice, up to 10 MB | One clip shows a visual or timing bug better than a paragraph, and it fits a Jira upload. You can start it from the right-click menu too. |
| Screenshot editor: arrow, box, pen, text, blur and crop | Point developers at the exact problem, and blur customer data before you share. |
| Passwords and tokens masked by default | Safe to attach to a shared ticket. No accidental credential leaks. |
| Follows new tabs, pop-ups, confirm dialogs and iframes | Payment and sign-in flows are recorded end to end. |
| Play it back, with closest match and pauses for password and OTP | Check the bug still happens before you file it, even after a small UI change. |
| Saved scenarios to play again after the fix | Retesting a fix is one click instead of redoing every step by hand. |
| A test that reproduces the bug: Playwright, Cypress or Selenium, one file or Page objects | The bug becomes a regression test for the team’s suite. |
| Copy bug, Copy screenshot, and a zip with the report, HAR and console | Pastes cleanly into Jira, Azure DevOps, GitHub or email. |
The free trial unlocks all of it. Pricing shows what stays free after the trial.
See and test the data behind every screen, without DevTools.
| What it does | How it helps you as a QA |
|---|---|
| Record every API while you use the site, grouped by screen | Know exactly which calls a screen makes, including ones the spec never documented. |
| Follows sub-domains, new tabs and frames, and ends by itself | Long journeys across apps are captured without watching the recording. |
| Health line: failed, slow, duplicate, repeated (N+1) and polling calls | Finds performance and backend bugs that the screen hides. |
| “Changed” badge when an answer’s shape changes since last time | Catches a backend change that breaks the front end before users do. |
| Test for one API: checks plus negatives (no auth, expired token, SQL and script text, missing or wrong-type fields) | A full positive and negative set in seconds, which you would otherwise write by hand. |
| GraphQL split by operation, with a check for answers that hide errors | GraphQL apps get readable lists and tests that fail when they should. |
| WebSocket and live-update capture, XML and SOAP answers | Chat, live feeds and older enterprise services are covered too. |
| Export the recording as HAR or Postman, tokens masked | Hand the traffic to developers or reuse it in Postman without editing secrets out. |
| Swagger: load all APIs from any Swagger page (old Springfox 2.x and Swagger 1.2 too), Execute, the whole spec as one zip | Contract tests for a whole service, straight from its docs page. |
| Playwright, Cypress, REST Assured and cURL, secrets as environment variables | Runs in CI without leaking tokens into Git. |
The free trial unlocks all of it. Pricing shows what stays free after the trial.
Find out if it copes with a crowd before the release does.
| What it does | How it helps you as a QA |
|---|---|
| Record a journey, get a JMeter .jmx | No JMeter scripting by hand. Start load testing the same day. |
| Load, Stress, Spike and Soak plans with pass and fail targets | The right test for the question: a normal day, the breaking point, a sale rush or a slow leak. |
| A k6 script.js as well | Teams on k6 get the same journey without switching tools. |
| users.csv with different data per user, plus a README | Realistic users from the first run, and anyone on the team can run it. |
| Quick check: response time and p95 for one API | A fast answer to “did this API get slower?” without a full load test. |
| Warns about calls that change data | No accidental 500 orders in a shared test environment. |
The free trial unlocks all of it. Pricing shows what stays free after the trial.
Catch everyday security slips and hand developers the fix.
| What it does | How it helps you as a QA |
|---|---|
| One-click passive check of the page, with a grade | A security baseline on every page you test, with no setup and no risk to the site. |
| Headers, CSP, cookies, CORS, mixed content and forms | Covers the common findings in a VAPT report before the auditors do. |
| Secrets, tokens and JWTs in the page, storage and URLs | Finds leaked keys and weak tokens that functional testing never looks at. |
| Debug leftovers: source maps, error pages, AI keys in page scripts | Release-day mistakes caught before go-live. |
| The fix for each finding, plus a regression test | Developers know what to change, and the test proves it stays fixed. |
| Report zip: report.html, findings and tests | Ready to attach to a ticket or a release sign-off. |
The free trial unlocks all of it. Pricing shows what stays free after the trial.
Test your AI chatbot without paying for an AI judge.
| What it does | How it helps you as a QA |
|---|---|
| Record a chat, get a test with checks for every answer | Regression tests for a bot whose answers change every time. |
| Must contain, must not contain, must refuse, answer time | Checks that still work when the AI words its answer differently. |
| Prompt injection and AI safety tests | Proves that “ignore your instructions” and role-play tricks don’t work. |
| Privacy tests: other customers’ data, card numbers, mixed-language tricks | Catches the leak before it becomes a headline. |
| Chat API failures: offline, error, empty and slow | Shows what users see when the AI service is down. |
| Chat box accessibility, links in answers, chat cleared after sign-out | Screen-reader users and shared computers are covered too. |
The free trial unlocks all of it. Pricing shows what stays free after the trial.
For manual testers starting automation, automation engineers who want to skip the boilerplate, and QA leads who want readable tests and reproducible bugs across the team.
Add TestMaadu to your browser.
No AI tokens needed. Free to try, nothing else to install, and your pages and test data stay in your browser.
See each module in action.
Watch a short video, or switch to Try it yourself and click through a live demo of the panel.
Point at an element, or list them all. Get the locator and the code.
Pick one, in four steps: click anything on the page, check the locator, choose an action and press Verify to see it work, then Copy code or Download it for Playwright, Cypress or Selenium. Find all: list the key elements on the screen and get a Page Object or a screen test for the ones you tick.
This is a live copy of the extension's panel. Click around, or press Play demo and watch it run.
- 1Select an element
- 2Check the locator
- 3Choose an action
- 4Verify it
- 5Get the code
- 6Find all
Sign in
Delivery
Recent orders
| Order | Status | Action |
|---|---|---|
| #1041 | Delivered | |
| #1042 | Shipped | |
| #1043 | Processing |
No AI tokens needed. TestMaadu's own engine writes the code. No API keys, no per-use cost.
Works in your browser. Your pages, elements, test data and bug reports never leave your computer.
Instant. No waiting on an AI service. The same element always gives the same code.
Signing in only stores your email and plan.No element selected yet. Click Select Element, then click anything on the page.
- 2Check the locator
- 3Choose an action
- 4Get the code
- Role
- Accessible name
- Locator strategy
Why this locator?
This element is no longer on the page. The code was built from the frozen picture, so you can still copy it.
Action
Framework
Playwright version
Language
+1 more in the extension: Java. Selenium 4 or Playwright Java, with a TestNG test class.
+2 more in the extension: Playwright · Java and Selenium · Java (Page Object and TestNG screen test).
Scans automatically when you open Find all and lists the key elements on this screen.
Scannedshop.example.com/account
+5 more elements in the extension. This demo shows 8 of the 13 it found here. Find all in the extension lists them all, 25 at a time.
Add to ChromeCopy, Page Object, Screen test and CSV are part of Lifetime access. One payment, no subscription.
Get lifetime access · ₹399One row per ticked element, with these columns:
- Name
- Type
- Group
- Status
- Playwright (JS/TS)
- Playwright (Python)
- Cypress
- Selenium (JS)
- Selenium (Python)
- In frame
Record the bug once. The report and the test write themselves.
Recreate the bug while TestMaadu records the steps, and a video with your voice if you like. Play the steps back to check it still happens, get a marked-up screenshot, the errors and a Playwright, Cypress or Selenium repro test, and save it as a scenario for after the fix.
This is a live copy of the extension's panel. Click around, or press Play demo and watch it run.
- 1Record
- 2Find the bug
- 3Play it back
- 4Generate the test
- 5Send the bug
- 6Save as scenario
Checkout
Order summary
Masala Chai × 1₹240
DeliveryFree
Gift wrap₹30
SPRING25−25%
Total₹240
Please agree to the terms to place your order.
Record the steps and a video, report it in one go.
Press Record, then use the page until you reach the bug. TestMaadu writes each step for you.
- 2Check the steps
- 3Add bug details
- 4Generate test code
- 5Send the bug
- 6Save as scenario optional
Free plan plays steps 1 to 5. Lifetime access plays them all.
Drag on the screenshot to draw a box, or blur private details.
Video
Click a step’s time to jump there. Copy bug can’t carry a video: use Download zip or Download video. The small testmaadu.in mark is on free and trial videos only.
Errors caughtEvidence
npx playwright test.Save as a scenario
Keep these steps to play them again later, after a fix or on another site.
Press Record and use your site, or open your Swagger page. Pick one API. Get the test.
Live Page APIs lists every API your site calls, grouped by screen, with reads, writes and failures for each. Select one, choose the checks and negative tests, and Generate. Export the recording as HAR or Postman. Secrets stay out of the code.
This is a live copy of the extension's panel. Click around, or press Play demo and watch it run.
Press Record and use your site. Every API it calls is listed by screen, right in the side panel. One API at a time, no DevTools needed.
- 1Record
- 2Select an API
- 3Choose what to check
- 4Negative tests
- 5Generate
Your cart
Your cart is empty. Add something from the Catalog.
Record reloads this page, then lists every API it calls while you move between screens and tabs of this site. Press End Recording when you are done.
Recording stopped. The APIs stay in the list until you clear them.
- 3Choose what to check
- 4Select negative tests
- 5Generate
Authorization becomes TESTMAADU_AUTHORIZATION, an environment variable. Your token never lands in the code.
+27 more negative scenarios in the extension Null, empty and 5,000-character values, script and SQL characters, wrong method, unknown address and more.
Framework
+3 more in the extension: Cypress, REST Assured (Java) and cURL.
Language
Swagger turns your Swagger page into API tests, one API at a time, in six steps: Load all APIs from this page with one click (or paste the spec link), Select an API, Try it with Execute for a real response, Choose what to check, Select negative tests and Generate. Need every API? Download all as one zip sits under the list. Works with every Swagger page: Swagger UI 2 to 5 (including the old green Springfox pages), Swagger 1.2 and OpenAPI 3, Redoc, RapiDoc, Scalar and Stoplight.
- 1Load the spec
- 2Select an API
- 3Try it
- 4Choose what to check
- 5Negative tests
- 6Generate
Example Petstore 1.0.0OAS 3.0
A made-up pet store API for trying Swagger. Sign-in: API key or Bearer token.
- Open your Swagger page in this browser tab.
- Click Load all APIs from this page. TestMaadu reads every API in it.
or paste the link
Spec needs a sign-in header?
Add a header name and value. It is used only to read the spec, sent only to the site you chose, and never saved.No Swagger file?
Load a Swagger file first. Its APIs are listed here.
Server: https://petstore.example.com/api/v3
BecomesBASE_URL in the tests.Download all 8 APIs as one zip
Language
+3 more in the extension: Java (REST Assured), Cypress and cURL.
- 3Try it (optional)
- 4Choose what to check
- 5Select negative tests
- 6Generate
Framework
+3 more in the extension: REST Assured (Java), Cypress and cURL.
Language
Generated locally in your browser. The spec and responses stay in the panel.
Execute sends a request only when you click it. It goes from your browser straight to the server you picked, never through TestMaadu. POST, PUT, PATCH and DELETE ask once more first, so run them against a test server. Your browser cookies are sent only if you tick “Send my browser cookies for this site”.
Record a user journey. Get a JMeter load test.
Use your site like a real user, press Stop, and download a JMeter or k6 plan with a load profile and pass/fail targets.
This is a live copy of the extension's panel. Click around, or press Play demo and watch it run.
Record a journey in the side panel and get a load test plan for many users.
- 1Record journey
- 2Use the site, then Stop
- 3Check the calls
- 4Pick a load profile
- 5Download the plan
Sign in
Signed in as qa.user@example.com
Find a product
Test typeJMeter load test
Press Record journey, then use the site like a real user. TestMaadu notes every page, form post and API call.
- 2Select the calls
- 3Add test data optional
- 4Set load and targets
- 5Generate
Load profile
Pass/fail targets
It adds items to a cart, many times over. Run it only on a test environment.
TestMaadu writes the plan. It never sends the load itself. You run the plan in Apache JMeter or k6, on your own machine or a load-test server, against a system you own or are allowed to test. The README in the zip has the install and run steps.
Quick check times one API call. In the DevTools tab, pick a captured call and choose Performance Testing: it always opens Quick check. See how long the call took and where the time went, set a time budget, and get a Playwright test that repeats the call and fails when it gets slow.
- 1Pick a call
- 2Performance Testing
- 3Set runs and budget
- 4Simulate a run
- 5Copy or Download
Captured calls · pick one
Pick a captured call on the left, then press Performance Testing.
- 2Set runs and time budget
- 3Generate
The test repeats it, so it could create orders many times. Run it on a test environment.
Check a page’s security. Prove the fix with a test.
Passive checks in nine areas, a grade from A to F, how to fix each finding, and tests that pass once it is fixed.
This is a live copy of the extension's panel. Click around, or press Play demo and watch it run.
- 1Check this page
- 2Review a finding
- 3How to fix
- 4Generate tests
- 5Download report
Sign in
build 2.3.1-dev · api: staging-api.example.com
GET /login · 200 · response headers
content-type: text/html; charset=utf-8strict-transport-security: max-age=31536000x-content-type-options: nosniffset-cookie: sid=••••••3f9a; Path=/; Secure; SameSite=Laxcontent-security-policymissingx-frame-optionsmissingreferrer-policymissing
Local storage
stripe_key: sk_live_••••••••a91cauth_token: eyJhbGciOiJIUzI1NiJ9.••••
Passive security checks for the page you are on: headers, CSP, cookies, CORS, mixed content, exposed secrets, risky forms, debug leftovers and JWTs.
- 2Review the findings
- 3Get tests and report
Checking shop.example.com/login
+1 more in the extension: Playwright · Java. SecurityTest.java with TestNG, all 26 check types.
What it is, and what it is not. The Security tab covers the repeatable, everyday security checks a tester can run on every page and every build. It does not replace a security tester or a full security review, and it only reports what it can see without attacking the site.
Chat with a website’s AI bot. Get the chatbot test.
Record real questions and answers, tick the checks each answer must pass, add negative tests, and get Playwright or Cypress code. No AI tokens, nothing uploaded.
This is a live copy of the extension's panel. Click around, or press Play demo and watch it run.
- 1Record
- 2Chat with the bot
- 3Stop
- 4Check answers
- 5Negative tests
- 6Generate
- 7Save
AI Chatbot
Chat with the bot,
get the chatbot test.
Press Record and ask the website’s AI chatbot your questions. TestMaadu captures every question and answer, then writes a Playwright or Cypress test that waits for the full answer.
- 2Check the answers
- 3Select negative tests optional
- 4Generate the test
- 5Save questions optional
Everything stays in your browser. No AI tokens, nothing uploaded. Emails and phone numbers are masked.
Captured so far
Only this tab is recorded. Quick-reply buttons you click in the chat count as questions.
Output
+1 more in the extension: Playwright · Java. A Maven project with every check and negative test.
Your free trial has ended. You can still record, review and preview the test; lifetime access unlocks copy and download.
Answers stay in your browser. Each test run asks the site’s real AI, which costs the site owner a little money.
What it checks, and what it does not. The AI Chatbot tab records real questions and answers and writes tests with plain text rules: the answer appears, mentions what it must, finishes in time, shows no personal data and refuses tricks. It does not score how correct or good an answer is; you decide that with the checks and Wrong answer. No AI tokens, and nothing is uploaded.
Capture once. Generate anywhere.
Inside the extension every tab goes step by step, with numbered steps. Underneath, the same three steps work for an element on a page, a bug, an API call or Swagger file, a user journey for a load test, a security check and a chat with an AI chatbot.
Select or record
Click an element on the page or list them all, record the steps to a bug (with a video and your voice if you like), record your site and select one of its APIs or load all APIs from your Swagger page, record a journey for a load test, press Check this page, or record a chat with the site’s AI chatbot.
Review and verify
TestMaadu shows the locator, the bug report, the request, the recorded calls, the response-time snapshot, the security findings or each chatbot answer with its checks. Choose the action, the checks, the load profile or the time budget, press Verify to run a step on the page, or Play it back to run a whole bug again.
Generate and copy
Get Playwright, Cypress or Selenium code, Page Objects, a bug report with a video and a repro test, a JMeter or k6 load test plan, a repeat-run performance test, security regression tests and a report, or AI chatbot tests with negative tests, with comments. Copy code or Download it.
In TypeScript, JavaScript, Python and Java. Load tests for JMeter or k6.
Questions, answered.
Does TestMaadu use AI, and do I pay for tokens?
No. TestMaadu builds code from what it finds on the page and in the request, using its own rules, inside your browser. There is no AI model to pay for, no API key to paste in and no usage meter.
Can TestMaadu test my website’s AI chatbot?
Yes, in the AI Chatbot tab. Press Record and chat with the bot the way a customer would. TestMaadu captures each question, the full answer (streamed replies too) and its timing, and fills in checks from each answer: no error, must contain, must not contain, answers in time, no personal data, and must refuse for trick questions. Pick negative tests such as a script tag, “ignore your instructions”, another customer’s details and a failing chat API, then get Playwright or Cypress tests.
It uses rules, not an AI judge, so it needs no AI tokens and gives no correctness score. When an answer is wrong, press Wrong answer and type what it should say, and the test fails until the bot says it. The chat stays in your browser.
Does my page or API data leave my browser?
No. Inspected pages, API requests and responses, performance timings, recorded journeys, recorded chats, security findings and generated code are never sent to us or anyone else. Only your sign-in (Google or email code) and your trial or lifetime status talk to our server.
Your API requests and responses are not saved to disk. A Live Page APIs recording is held only in temporary browser memory, so it survives closing the side panel, and is cleared when you press Clear or close the browser. Export writes a HAR or Postman file only when you click it, with tokens and passwords masked. Two small things stay on your device only: Swagger remembers the last spec link and server, and Live Page APIs keeps the shape of each API’s answer (field names and types, never values) so it can tell you when it changes. Log a Bug scenarios are kept only when you press Save scenario. The Privacy Policy has the details.
Which frameworks and languages does it support?
Find Element generates Playwright code in TypeScript, JavaScript, Python and Java, Cypress code in TypeScript and JavaScript, and Selenium code in JavaScript, Python and Java (Selenium 4, with a TestNG test class). Switching keeps the element selected and rebuilds the code, and a Playwright version option (Any or 1.64+) gives shorter locators for Playwright 1.64 and later. Find all exports a Page Object or a screen test in the same languages, Java included. The Log a Bug test code is Playwright or Cypress (TypeScript or JavaScript) or Selenium (Java with TestNG, or Python with pytest), as one file or as a Page objects project. API Tests (Live Page APIs and Swagger) generate Playwright tests in TypeScript, JavaScript and Python, and also Cypress, Java (REST Assured) and cURL. Quick check generates Playwright tests in TypeScript, JavaScript and Python. The JMeter load test is written for JMeter (.jmx, built-in elements only, no plugins) or k6 (JavaScript). Security regression tests and AI Chatbot tests are Playwright (TypeScript, JavaScript, Python or Java) or Cypress (TypeScript or JavaScript).
Will it work on any website?
On any site you can open in Chrome or Edge, including your own staging and local apps. Browsers do not let extensions read a few pages, such as the Chrome Web Store, Edge Add-ons and chrome:// or edge:// pages, so those cannot be inspected.
Does the Security tab attack my site? Does it replace a security tester?
No to both. The checks are passive: TestMaadu only reads the page, its cookies and its response headers. The headers come from DevTools when it is open, otherwise from one re-request of the page, which is skipped for logout, delete, confirm, reset and OAuth sign-in links, and for URLs that carry tokens or one-time codes. Nothing is attacked.
It covers the repeatable, everyday checks you can run on every page and every build. It does not replace a security tester or a full security review. Run the generated tests only against applications you own or are authorised to test.
How does the free trial work?
Install TestMaadu and sign in with Google, or with any email and a 6-digit code, to get 3 days of full access to every tab, with Copy and Download. No card needed. Before you sign in, the free features already work and generated code shows as a preview. The trial belongs to your account, so reinstalling the extension or switching to another browser continues the same trial instead of starting a new one.
What happens when my trial ends?
Each tab shows “Your trial has expired. Free features still work.” with a Get lifetime access · ₹399 button, and the header pill says Trial expired. Generated code is shown only in part and copying is turned off until you unlock lifetime access. The locator stays free to copy for 3 selected elements a day, Find all shows its list with the locators hidden, Log a Bug keeps working with its free features (the video needs lifetime access), Live Page APIs still lists a page’s APIs, Swagger still loads, selects and executes, the Quick check snapshot stays visible, the 1-user Smoke load test plan stays free, Security keeps working with its free features, and AI Chatbot still records and reviews chats with their checks (its test code stays a locked preview). Your account and settings stay as they were.
Get lifetime access opens one window that lists what each tab unlocks. You pay ₹399 once in that same window, and an “Enjoy Maadi!” receipt confirms it.
Have 75 seconds?
Watch the full product demo: Find Element, Log a Bug, API tests, performance and security, all in one browser extension.
75-second product demo
Skip one weekend movie.
Get automation code for life.
Try it free for 3 days: just sign in with Google or your email. Then pay ₹399 once. No AI tokens needed, ever.
Every feature, free. Sign in with Google or your email to start. No card needed.
Add to ChromeNo AI tokens needed. No subscription. Every future update included.
Get lifetime access