LARRUZ demos Dscryb, an AI that summarizes videos by combining speech transcription with frame-by-frame visual analysis. It catches details like a disappearing shirt, a tattoo, and a music production setup that text-only summaries would miss.
Dscryb is a huge step up for AI generated summaries.
It takes video as an input and then it uses AI to transcribe the content of the video, summarize that. But it also uses vision to look at the frames and then take the information from that and then includes it in the summary. So you get an audio visual summary of the video. Obviously it works on the examples here, but let me show you one of mine too.
For example, and my shirt is gone.
So this is just a random clip from a previous video we created where I show a camera transition where this over shirt disappears. It vanished. Let's see if it summarizes it.
Well, alright then after 24 seconds we have a result here. Look at this. In this video, a man with long hair and a beard is captured in a personal space dedicated to music production. Yeah, close, huh? Displaying his passion for the craft. But look at this part. A distinctive tattoo on his arm and his casual attire, including a moment where his shirt seems to disappear, adds a relaxed yet focused atmosphere to the setting. So as you can see, it really picked up all the action, including the little detail off the tattoo back here.
