ROHKI Intersection

We moved to Japan for 6 months to make this episode. We flew out five talented AI artists from around the world so that we could all collaborate in the same room together. In terms of the tools used we pretty much used every AI tool available to make this. The main workflow was filming ourselves acting out the scenes on a green screen and putting that through Viggle to get it looking close to our characters. We then ran that through a custom workflow we made in ComfyUI along with custom trained LORAs for each character that we made using midjourney. We had to do separate passes for the body and face to maintain high resolution in the face. Our 3D artist made all the backgrounds in 3D so the next step was compositing the characters into the 3D backgrounds using Davinci Resolve. Certain shots were able to be done with Luma or Runway image to video. We also used Topaz and Magnific to upres all the assets. Finally we made all the music and sound effects ourselves and did all the voice acting. This was such a passion project for us and we hope you enjoy