Artificial Intelligence in Playwright - No more locators, just natural language 🦾⚡️

[@andreaskarz]

I'm curious to see how this works with really complex websites. And above all, how reliably the tests run, because a big advantage of Playwright is the reliable tests with little flakiness. And what I also see as a problem is that the test results should always work with static data and static results. Only then are they truly reproducible.

[@ladislavzahnas7618]

I am a bit skeptical how this will work for big complex websites. But what I consider the worst...I have no idea where are you sending your information. So proceed with caution and consult with your security before you try to implement it in production.

[@ExecuteAutomation]

I agree on with you mate, I will check with the Authors and get back to you on this.

[@abesoky]

Excellent information! Thanks for sharing, very clear

[@amrsalem6655]

Very interesting demonstration, could i ask you whats the recording tool you are using ?

[@queenstownswords]

Hey Karthik. Good video. It needs a C# binding so we can leverage it with specflow... or take it one more step and use the BDD statement to drive the AI test!

[@ExecuteAutomation]

True mate, so far, this ai() function is implemented in TS/JS, hopefully soon in other languages too.

[@tuberklz]

so everytime i start the same test i will be using that 8 ai api call for just one same test scenario. maybe it would be better if output was clear playwright test code

[@nikolamilevski6253]

Test is only running i headless mode, its not opening chrome browser and i have config the --headed in package.json

[@JayaSuryaBoodaraju]

Will this work for js files aswell or is it limited to the ts

[@gdd546]

dont really work, telling it to click on sign in link but its clicking on a community link and passing the test. tried it on the new era caps website

[@soumyasidharthamohanty6379]

Hay Karthik , Can you please let us know where to define token. and how you have defined it in your bash profile. And are making any call of token in framework or not .

[@ExecuteAutomation]

All you need is to run this command by replacing it with your ZeroStep token $ export ZEROSTEP_TOKEN=""

[@RodrigoASouza-hd1og]

Hi I cant select a item on list dropdown

[@prathyushareddy1185]

Hello, I was doing POC on Playwright with Zerostep AI and got an error like 'An unknown error occurred' while executing in Chromium. could you help me how to fix the issue.

[@ExecuteAutomation]

This should be due to the issue I mentioned on the video, try again and it may work

[@prathyushareddy1185]

Please let me know how to connect with you. bcz still am seeing the issue.

[@drmrgtx]

We still cannot use the code generator for complex page to find the correct page object… so why anyone thinks AI could help to build valid code for a highly structured, dynamic HTML

[@NerdyThrowbackTech]

Looks promising

[@ExecuteAutomation]

True ! More videos are coming soon on the same.

[@ruksharalam173]

Would be great if we get something similar for Python Playwright.

[@ExecuteAutomation]

I don't think the implementation is far away for this

[@LuisBorges0]

Better change the OpenAI API key, okay?

[@oscargustavorodriguez3123]

I'm scared

[@trilogen]

Will never trust Cloud run AI with my data. Locally run AI like LMStudio will gain more popularity over the years.