{{g.name}}
{{g.platform}}
One 18-screen regression, written in Dev Scenario and as native XCUITest and Espresso suites tuned for speed. Same screens, same taps, same checks, same order.
Dev Scenario is fastest on every app
{{g.name}}
{{g.platform}}
Seconds for screens 2–18 in one session, shorter is faster; bars are scaled within each app (Android: screens 2–18 without the modal code sheet, which it hasn't been measured on yet). Dev Scenario uses its in-app SDK. Maestro {{maestroVersion}} runs the same steps with its own commands; its totals leave out Pinch, as Maestro has no pinch gesture, so they cover one screen fewer than the others. Maestro on the Compose Multiplatform app is not measured yet. {{setup}}. Faster of two runs each, after one warm-up run.
Real time, no cuts: the Android sample regression on an emulator with the in-app SDK, all 18 units in 15.2 s. A single run, not the chart's measurement: it also includes the launch and the code sheet, which the chart leaves out.
{{screens}}
screens, timed one by one below
{{iosSteps}}
steps ({{androidSteps}} on Android)
{{actions}}
taps, typing, swipes, drags
{{iosChecks}}
checks ({{androidChecks}} on Android)
1
app launch, no restarts
Bold marks the faster tool on each screen. Seconds: the faster of two runs per suite; screen times are rounded, totals are exact.
Native tests tuned for speed
Each native suite uses its own tool's fastest APIs and waits: Compose test clock and idle sync on Android, XCUITest idle sync and short gestures on iOS. No sleeps.
Same coverage, step for step
Every native test mirrors one Dev Scenario screen: the same actions, the same assertions, in the same order.
Launch left out
Native runners start the app outside their own timer, so every tool is timed on screens 2–18.
Wall-clock time
From the first timed step to the last, as each tool reports it. App install and device boot are not counted. On iOS the system file picker is reset to its default place before every Dev Scenario SDK and native run.
A project's own native UI tests, run as its developers wrote them, against the same tests written as Dev Scenario flows with the in-app SDK.
{{p.kind}}
{{p.speedup}}{{p.testsLine}}
Their tests, unchanged
The Espresso suite runs with the project's own CI command, at commit {{p.commitShort}}.
Same setup per test
Each flow starts with appAction: the app runs the same reset as the test's @Before and opens the same screen. The app's debug build registers these reset actions.
Seconds from the first test to the last, lower is faster; the faster of two runs each ({{p.runs}}). Unlike the screen-by-screen benchmark above, Dev Scenario's time here includes its one app launch. {{p.setup}}.
The in-app SDK is the fastest. Without it, Dev Scenario drives any build from outside, with WDA on iOS or UIAutomator on Android.
Seconds. From-outside times are earlier single runs, measured before the latest in-app SDK runs (iOS: on the 17-screen regression, before the modal code sheet screen was added).