This study demonstrates the use of a neural interface system to enable people with tetraplegia to control a robotic arm for three-dimensional reach and grasp movements. Two participants, S3 and T2, who had long-standing tetraplegia due to brainstem stroke, were able to control the robotic arm without explicit training. The arm was controlled using signals decoded from a 96-channel microelectrode array implanted in their motor cortex. S3, who had the sensor implanted 5 years earlier, also used the robotic arm to drink coffee from a bottle. The results show that even years after injury to the central nervous system, people with tetraplegia can use a small sample of neural signals to perform complex tasks, suggesting the feasibility of restoring lost arm function through cortically driven commands.This study demonstrates the use of a neural interface system to enable people with tetraplegia to control a robotic arm for three-dimensional reach and grasp movements. Two participants, S3 and T2, who had long-standing tetraplegia due to brainstem stroke, were able to control the robotic arm without explicit training. The arm was controlled using signals decoded from a 96-channel microelectrode array implanted in their motor cortex. S3, who had the sensor implanted 5 years earlier, also used the robotic arm to drink coffee from a bottle. The results show that even years after injury to the central nervous system, people with tetraplegia can use a small sample of neural signals to perform complex tasks, suggesting the feasibility of restoring lost arm function through cortically driven commands.