Stream detection in realtime.
Casinolytics computer vision identifies games, their providers, and casino branding on tracked broadcasts. Captured frames connect those detections to the stream data you can review.
From detection to airtime.
Computer vision reads every captured frame and names the game on screen. Lobbies, menus, and transitions read “Untracked”.
Casino branding is detected separately, across every captured minute.
An explanatory reading of the scale, not measured results from the footage above.
Detections across a broadcast. Minute by minute, over the live viewer curve.
Open any tracked broadcast and see what was on screen, and when.
Stream Summary
Tracked broadcast / Sep 4, 2026
- Airtime
- 90 min
- Casino Airtime
- 76 min
- AI Coverage
- 93%
- Peak CCV
- 4,110
- Avg CCV
- 3,043
- Viewer Hours
- 4,565
Viewers per minute
Concurrent viewers across the broadcast
Check any minute against the stream. Open the capture behind any detection.
- No fabricated zeros: a minute the model could not read is “Untracked”, and a figure the count could not reach is a dash.
- No “views”: the platforms only expose concurrent viewers, so that is the only audience figure you will see.
- Every abbreviated number reveals its exact total on hover.
- Curated images training our computer vision models
- 3.2M+
- Computer vision models on the watch
- Since 2020
Casino branding, read off the screen.
Casino branding (logos, symbols, and colours) is its own detection class, measured across every captured minute, and an operator enters the rankings only above a minimum exposure floor.
- Observation
- On-screen casino branding
- Measured across
- All captured minutes
- Market ranking
- Minimum exposure required
Every detection becomes data. Yours through the API.
Airtime, viewer minutes, and peak viewers for studios, games, streamers, and casinos, through the Casinolytics API. Available by request.
Stake casino branding
Drac's Stacks game
Massive Studios provider