SmartTile reads the Verkada Command API and puts every camera, intercom and sensor on the same pane as your circuits and switches: imported, polled every minute, ticketed when one drops. Your people see the cameras SmartChoice assigns them, and nothing else.
A camera that goes dark is usually noticed when someone needs the footage. Verkada Command knows the moment a camera stops reporting, but that knowledge stays inside Command. SmartTile asks Command every minute and treats a Verkada camera exactly like a switch or a firewall: it goes red, the outage is recorded, the ticket opens, and it closes on recovery.
Cameras, intercoms, environmental sensors, alarm devices and chimes are read from Command and written as devices with model, serial number, MAC, local IP and firmware. Each Verkada site is matched to a SmartTile location by name. A re-import updates, never duplicates.
A background poller reads Command's online state for every imported device (on by default, interval 1–120 min). A status is written only on a real change, and every change lands in state history, so uptime and reports count the outage.
A device that drops out of an otherwise full listing goes DOWN after about 15 minutes. An empty or failed listing never marks anything down; after three blind polls the UP devices turn grey (UNKNOWN) and the NOC is e-mailed.
Verkada devices go through the same alert engine as ping and SNMP devices: outage e-mail, ticket in your ITSM system after the open delay, close on recovery. Incidents ticket too: a plate-of-interest hit opens one ticket in Rev.io and your ITSM system with camera, location, time and a 48-hour clip link — on by default, switchable per company.
Latest thumbnails and a live HLS stream per camera from Command's footage API, inside SmartTile. Recorded playback is available per camera when it has been granted.
Door unlock (3–30 s pulse), door schedule overrides, site lockdown with a one-hour auto-end, intercom answer / deny / unlock, and Verkada access users: create, deactivate, cards, PINs and access groups. Every action needs a written reason.
| Role | What works |
|---|---|
| SmartChoice NOC (super admin) | The Verkada console: inventory, live view, events, alarms, environment, reports, import, Customer Access. |
| Physical security tech | Door unlock and intercom call handling. |
| Physical security admin | Site lockdown start / end, door schedule overrides, access-user and credential changes, audit-chain check. |
| Customer users | My Cameras: only the cameras granted to their company, live, and recorded only where allowed. |
/tokenexchange the org API key for an API token (cached 28 min)/cameras/v1/devicesinventory and online state — import and every status poll/access/v1/doorsdoors, their sites and paired cameras/events/v1/accessbadge and door events — live stream, badge in/out reports/cameras/v1/footage/thumbnails/latestlatest snapshot per camera/cameras/v1/footage/tokenshort-lived token for the live / recorded HLS stream/access/v1/door/{id}/unlockdoor unlock — reason required, logged/access/v1/sites/{id}/lockdown/start · /endsite lockdown — admin role, auto-ends after one hour/access/v1/access_users…access users, credentials and access groups/cameras/v1/analytics/lpr/license_plate_of_interestplate-of-interest reads, every 30 s/core/v1/audit_logCommand's own audit log, merged hourlyMy Cameras intersects the live inventory with the grants for the caller's company on the server. No site names, addresses, serials, IPs or MACs, and never a Command login. Off per company by default, with a platform-wide switch.
Every unlock, lockdown, schedule override and access-user change is written to a hash-chained log per company, with user, time, reason and result. The chain can be verified end to end.
The Verkada API key lives in the server environment, not in the database, and is never sent to a browser. Streams use Verkada's short-lived footage token.
The status poller has its own on/off, interval, missing-device and blind thresholds (each with ⓘ help), plus a global Verkada kill switch. Incident tickets have a per-company switch on the Incidents page and a platform switch.
| Objection | Answer |
|---|---|
| “Verkada already tells us a camera is offline.” | Inside Command. SmartTile turns the same fact into a ticket in the system your team works from, next to the circuit and switch at that site, and counts it in uptime. |
| “A vendor API can just drop a device.” | SmartTile treats a device missing from a full listing as DOWN after about 15 minutes, and never marks the fleet down on an empty answer. |
| “Can our staff see other companies' cameras?” | No. The customer page is filtered on the server by your company's grants; anything not granted does not exist for your users. |
| “Could an incident ticket land on the wrong customer?” | No. The company comes from the camera's grant or imported device, never from the Verkada event. If SmartTile cannot tell, the incident is recorded and no ticket opens — and it says why. |