We currently use WebdriverIO for our web testing, but we’re actively considering a migration to Playwright. One concept that caught our attention is self-healing test automation — tools or frameworks that can automatically update or adapt locators when the UI changes.
We’re curious to know:
Have you implemented self-healing in your test automation setup?
What tools or plugins have you used (especially with WebdriverIO or Playwright)?
Did it help reduce flaky tests or just add more complexity?
In your experience, is it worth integrating, or is strong locator strategy and test design enough?
We’d really appreciate hearing your real-world experiences — good, bad, or surprising. Any recommendations (or warnings) would help us make an informed decision as we evaluate our next steps.
I find myself looting for AI-inspired post here in MoT, and stumbled upon this article:
It might help to see what the community say about this in the following thread too:
Honestly reading the article and reading this thread still make me curious if these “self-healing” AI-assisted frameworks can do more than just locator change
I wonder if there is an update at present-day where it has been proven to be super useful… and if it has become really-really “self-healing”?
So, I haven’t used full self-healing quite yet, but I’ve looked into it a bit while exploring WebdriverIO and Playwright. Few thoughts:
Locator strategy stands above self-healing; more often than not, a strong locator strategy (think: data-test-ids, stable attributes, never-brittle selectors) will cause less grief than relying on auto-healing.
Tools I’ve seen: Healenium (in the Java/Selenium world) tends to come up a lot, but for the Playwright world, I’ve usually seen folks trying and experimenting with some community plugins or just writing their own locator retry logic.
Pros and Cons. Self-healing sounds magical, but it ends up being complex most of the times. Debugging gets tricky when the framework silently “fixes” locators - sometimes you’d rather know something broke!
That being said, I think lightweight healing techniques think fallback locators or intelligent retries come into play whenever the UI undergoes constant changes. For a big team or fast-paced product, the trade-off is worthy.
If you’re migrating to Playwright, my gut says: first go strong with locator strategy and test design, and then lead with self-healing only if scaling flaky tests become a pain.
How about the community in here, has anyone found a good solution with, say, WebdriverIO or Playwright, and do you think this aspect is still evolving?
