01
Capture and fingerprint
Live audio is captured locally and converted into compact fingerprints directly on the television.
Edge recognition platform
EdgeACR
On-device recognition for the broadest television fleets
EdgeACR validates ad exposure locally, distributes compact catalogue deltas, and returns privacy-safe matched events across legacy, constrained, and mixed television environments.
Architecture
How EdgeACR works
Three stages, all running at the edge: fingerprint locally, stay current with compact catalogue updates, and return only matched events.
- 02
Sync compact catalogues
Small catalogue deltas are delivered to the device, so recognition stays current without heavy refresh cycles.
- 03
Return matched signals
Only validated matched events move outward, encrypted and ready for analytics, planning and activation workflows.
Why EdgeACR
Built for fleets other systems leave behind
Runs on constrained fleets
Built for older and lower-spec hardware profiles, not just flagship devices.
Keeps media on device
Recognition happens locally rather than streaming raw media outward.
Distributes only what changes
Catalogue delivery is incremental and compact, keeping bandwidth low.
Produces usable commercial signals
Matched events arrive ready for analytics and activation workflows.
Request a demo
See EdgeACR in a live workflow
Fleet-scale on-device recognition, compact catalogue delivery, and privacy-safe signal output.