Raypack® is a registered trademark and a compression, playback and delivery solution that uses Artificial Intelligence based on recurrent Neural Networks. We expand the transparent compression layer using trained network models whether our machined models provide better results for any compression type like for image recognition, object detection, optical flow and motion estimation detection. Furthermore, we are releasing our compression models via NVIDIA Cuda Digest so that we can extend more better results with the trained data.


The Raypack transcoder is able to encode 360 Videos from Equirectangular 360 Videos to Cubemap 360. Cubemap 360 reduces the filesize up to 30% against  Equirectangular 360 videos.


Our output formats for 360 are HLS with Multiple qualities profiles.

RAYPACK® Video Compression Demo
This demo demonstrate Raypack compression engine in combination with our native HTML(5) videoplayer. The 100KB Player detects automaticly the client network condition and choose on the fly the correct profile so that there is no buffering or delay at all.