Project status

Hi All,
Is this project still under development? Or are there recommended forks?
I'm wondering if there is planned support for video streams and paralled prediction?
At the moment prediction fails when multiple images are send in short succession.