Run code in a session
exec runs WebdriverIO code in the open session. Use it when a step is more than a single click or fill, and for every assertion.
npx wdio session exec -e "await browser.getTitle()"
npx wdio session <<'JS'
await $('aria/Cart (1)').waitForDisplayed()
JS
Always await commands. $ returns one element and throws when it is missing. $$ returns a list. There is no sync mode and no browser.element.
Names you declare stay available in the next exec. A top-level import is loaded from the project directory.
Assertions​
Put assertions in exec with expect-webdriverio:
npx wdio session exec -e "await expect($('h1')).toHaveText('Cart')"
Use visual check <tag> when the question is how the screen looks. That command needs @wdio/visual-service:
npx wdio session visual check cart
visual accept cart copies the latest actual image for that tag over the baseline. It does not copy older images that share the tag prefix.
When to use a shortcut instead​
click, fill, type, press and tap are shorter than exec for one interaction, and they print the WebdriverIO line they ran. Prefer those with a ref from the latest snapshot. Use exec for waits, assertions and anything that needs more than one command.
Next steps​
- Export a test — save the steps, including
exec - Commands —
execandvisualflags
