Popped out on its own so it can sit on a second monitor during a show, independent of the main docs page. Everything here works exactly like the embedded version — same buttons, same shortcuts, same log capture.
Not connected.
Two devices at once: add each via Add Device. With Auto-connect on, both reopen by themselves next time. Mark (or M) drops a numbered marker the instant you press the physical button, so both devices’ reactions can be read against the same moment. Click a device’s chip to mute it; the ✕ closes it.
Capturing starts on connect — nothing to configure. The last ~4,000 lines are kept and survive a disconnect, so you can still Copy, Download or Report after unplugging. Every line is matched against ~80 error patterns taken from the devices’ own firmware, and hits collect as clickable chips so you don’t have to read the whole log. Report Issue on GitHub pre-fills a bug report; nothing is sent until you click Submit there, since a log can contain your WiFi network name. Copy report puts the same report on the clipboard for an email instead.
Open .log (or drop a file on the console) replays a saved log through the same identification and issue detection, in any browser. prev / next (or [ and ]) walk the flagged lines in order; the Flagged / Errors / Warnings / Notes pills show only those lines; / jumps to the filter box.