Guides / Vision language model
A vision language model (VLM) for live video
Edge Video AI is a vision language model (VLM) that watches live video, reads what's on screen and what's said, and turns it into signals ads and shopping can act on in the moment.
Works alongside your existing playout and ad server.
The model
Sceney, the vision language model from Edge Video AI
Sceney, the vision language model from Edge Video AI Apps built on Sceney can be switched on over a live stream, or built by typing what you want.
Live enrichment turns what the model reads into IAB categories, scene keywords and brands seen on screen. The first two uses are ads and shopping. Autonomous Commerce for ads adds that context to the publisher's existing ad request, working alongside their playout and ad server. Shopping turns the moment into something you can buy.
Connected TV is where we start. The market is all video. Edge Video AI was founded in 2019, in The Hague.
Two ways to a screen
On a stream, no TV software at all. On a TV, Edge ACR.
Read the live stream itself
If Edge can see the stream, it can read it, on any screen that stream reaches. Nothing has to run on the viewer's television.
Edge ACR on the TV
Edge ACR is our own content recognition on the TV, talking to the vision language model (VLM). It works on apps, HDMI and free-to-air. The picture never leaves the set.
What one read carries
Signals ads and shopping can act on.
Each moment becomes IAB categories, scene keywords and brands seen on screen, in shapes that ad systems and software already understand. The model reads the picture, what is said, and the text and logos on screen. It describes the programme, not the person: it does not identify or profile viewers.
Edge can add information to an ad request; it cannot block a request. Works alongside your existing playout and ad server.
Go deeper
Sceney
The vision language model from Edge Video AI, and the apps built on it.
How the model works
What one read returns, the two paths to a screen, and the research behind it.
IAB categories in live ad requests
How live enrichment reaches the ad request.
Autonomous Commerce for ads
Adds that context to the publisher's existing ad request.
Questions
- What is a vision language model (VLM) for live video?
- Edge Video AI is a vision language model (VLM) that watches live video, reads what's on screen and what's said, and turns it into signals ads and shopping can act on in the moment.
- What is Sceney?
- Sceney, the vision language model from Edge Video AI Apps built on Sceney can be switched on over a live stream, or built by typing what you want.
- What is live enrichment?
- Live enrichment turns that into IAB categories, scene keywords and brands seen on screen.
- What is Edge ACR?
- Edge ACR is our own content recognition on the TV, talking to the vision language model (VLM). It works on apps, HDMI and free-to-air. The picture never leaves the set.
- What are the first two uses?
- The first two uses are ads and shopping. Autonomous Commerce for ads adds that context to the publisher's existing ad request, working alongside their playout and ad server. Shopping turns the moment into something you can buy.
Start with Sceney
Sceney, the vision language model from Edge Video AI