The initial input images for the clips are made with an SD 1.5 mix in A1111 and then animated in Runwayml gen3 I2V and finally extra facial animations added with Live Portrait V2V in comfyUI. Pleasantly surprised how well the extra facial animations fit onto the character.