Score is an open, permissionless computer vision network that operates as Subnet 44 on Bittensor. Video is split into short clips and distributed to a network of miners, who run computer vision models locally and return structured annotations that identify and track objects frame by frame. Validators verify those submissions through a combination of frame filtering, keypoint validation, and geometric checks, and miners are rewarded according to the accuracy and speed of their output. Score's initial focus is Game State Recognition in football, and the framework is sport-agnostic by design. The team behind Score also builds Manako, a commercial vision product that runs on Subnet 44. SN44 is the subnet's alpha token under Bittensor's dynamic TAO (dTAO) model. It is used to stake toward Subnet 44 and to distribute emissions to the miners and validators securing the network.