Hello!
When Webdriver BiDi becomes available officially (non-draft releases), it will be evaluated for any potential uses within SeleniumBase.

Are you aware that SeleniumBase has a CDP Mode that uses the CDP API directly without using Selenium?

Can you explain how you compared Playwright with SeleniumBase? Based on the numbers I've seen through various analytics, both frameworks run at approximately the same speed when used with optimal settings. I've compared that data during various live demos and public presentations. I'd be curious to see how you made your calculations and reached your conclusions, as the data I've seen shows something quite different from what you're reporting.

Thanks for your support!

Hi Michael,
Thanks for your time and reply!

Yes, we're aware of the CDP mode of SeleniumBase, and we're actually looking into it to check if it can help reducing the execution time of some test scenarios which are taking lot of time (so in our case, not intended to be used for anti-bot detection).
However, CDP mode is not a perfect option for us, as it is only compatible with Chrome/Chromium and not with Firefox (unless I'm wrong?) + we need W3C compliant, standardized API like WebDriver is providing.

About the comparison between Playwright and SeleniumBase, this is really interesting that you have analytics that proves that both frameworks are running at ~ the same speed! That could indicate, indeed, that we may miss some optimal settings.
Would you mind to list these settings optimal settings for SB?

On our side, we did our comparison by writing the same test scenario on a Web application that we are testing, one with SeleniumBase, the other one with Playwright (https://github.com/microsoft/playwright-pytest).
Both scripts were doing the same clicks and asserts, on Chrome (headed) and were built on top of Python + Pytest.

We saw that Playwright was taking 2x less time to complete the actions, and we saw that Playwrite seems to less wait for the full page to be loaded, performing actions more quickly.
Based on that, we also tried different load strategy / JSwait of SeleniumBase: while it was better, it did not compete with Playwrite.

FYI, our application is using old technologies/structure, such as lot of JS and iFrames - sometimes nested, I do not know if this can explain some latency.

Since you have already compared both frameworks, we will compare again both frameworks but this time on an online Web page (one of the Web page you are proving in the documentation for testing purpose).
We will also pay attention to the settings and we will also be able to provide the POC as needed.

Again, I'll be really interested of the optimal settings of SB you were referencing, if you could provide those.

Thank you again for your time and support!