Jump straight to the step that broke
Record any automated run and mark where each test begins, and the video comes back as one clip per test — a scrubbable timeline of steps with the failure marked, so a red build opens on the exact frame it broke.
Free device-minutes to start · no credit card required
A red build names the test. It doesn't show the frame.
Record the run and every test becomes its own clip, named the way your suite named it. The list shows which one went red — open it and you land on the moment it broke, not somewhere in a forty-minute file.
The run, on a timeline
A clip per test
Each test gets its own video, labelled with the name your suite already uses.
Open on the failure
The failed clip stands out in the list — one click and you are on the frame it broke on.
Pass or fail, at a glance
Every clip carries its result, so you skip the ones that passed.
One line in your hook
Mark test boundaries once in the before-each hook you already run. Nothing else changes.
Android splits itself
Appium runs on real Android phones are cut into clips automatically — nothing to add.
Kept, then yours to keep
Recordings stay 30 days on every plan; download anything you want for longer.
Send the clip, not a paragraph
Share a run by link with anyone who needs to see it — no account required — and revoke the link when you are done. The developer opens the failing clip and watches it happen, instead of trying to reproduce your description.
Everywhere your tests already run
The same recording on every machine and device on the fleet.
Real browsers
Chrome, Edge and Firefox on real macOS and Windows machines, plus Safari on macOS.
Real phones
Appium runs on physical Android handsets and iPhones record exactly the same way.
On when you ask
Turn it on per run in your capabilities — off by default, so you never pay to film a green suite.
Frequently asked questions
Ask for it in your capabilities: set "recording" to true inside testkase:options. It is off by default, so a run that does not need video does not pay for one.
Almost always the value was quoted. "recording": "true" is a string, not a boolean, and the run will complete with no video at all. It is an easy thing to get wrong in a YAML pipeline, so it is worth checking before you rely on the recording being there.
Tell us where each test starts. In Playwright that is a single page.evaluate carrying a testkase_action with setTestName; in Selenium it is the equivalent executeScript call. Put it in the hook your suite already runs before each test. Appium runs on Android need no call at all — they are split automatically. iOS is the exception: its recording comes back from the device as one blob at the very end, with nothing to cut, so an iOS run gives you a single continuous video rather than clips.
No — you get one or the other. A run with test markers produces the per-test clips instead of a single continuous file, not in addition to it. If you want one video of everything, leave the markers out.
Yes. A matching setTestStatus call records the result, so a failure stands out in the list without you opening each clip.
All of them — Chrome, Edge and Firefox on macOS and Windows machines, Safari on macOS, and Appium runs on real Android phones and iPhones. Per-test clips are the narrower set: every browser can be split, Android can be split, and iOS cannot yet.
30 days, on every plan including free. Download anything you need to keep for longer.
Yes. You can send a link to a run and revoke it later. Bear in mind the person holding it sees the whole session, including the network capture, so think before sending one outside your team.
Stop asking people to reproduce it
Send them the clip of the test that failed.