Overview
What this organization has, what it used this period, and what it did most recently. Every figure below names the endpoint it came from — and the ones that have no honest source are not here at all.
Complete and attributable — every value on screen comes from a named field of a receipt.
Devices and leases
GET /v1/devices6visible
3reachable
- 3 Online
- 1 Offline
- 1 Unauthorized
- 1 Unknown
How many are in use is not shown. No browser-reachable endpoint reports lease state, and reachable is not the same as free: a reachable device may be held by another run right now. A zero here would say the fleet is idle, which nothing here knows.
This period's usage
GET /v1/usage7h 46m
27,977,800 ms across 71 settled leases on 4 devices.
September 2026 — read from the database at 2026-09-03T14:22:31.884Z
This period's cost
GET /v1/usage469minutes reported
Echoed from the billing reporter exactly as reported. This console does not compute billable minutes — the rounding rule has one owner, and a second implementation would be a second source of truth for money.
Prices are not shown here. They live on the website; a copy of a price is a second price.
Recent actions
GET /v1/receiptstypeTextdev_pixel7_a192,900 msReceipt status: Needs reviewpressKeydev_robotarm_r044,600 msReceipt status: Needs reviewtapdev_pixel7_a191,300 msReceipt status: Failedswipedev_cloudphone_c021,100 msReceipt status: FailedopenAppdev_emulator_e01800 msReceipt status: Failedtapdev_pixel7_a19882 msReceipt status: Committed
There is no latency figure on this screen. The straightforward aggregate over action receipts returns a number with an impeccable source that is wrong by two orders of magnitude — it counts synthetic probe rows — and boot time has no instrumentation at all. Per-action timings, unaggregated, are on each device's page.