Soft mechanical sensors and actuators have revolutionized fields such as health monitoring, human-machine interaction, and soft robotics. However, existing fabrication methods, including 3D printing, molding and casting, lithography, and particle flow spinning, face challenges in producing highly integrated architectures, achieving multi-functionality, and often result in low aspect ratios. In this paper, we introduce thermal drawing as a materials and processing platform to develop advanced soft, multi-material magnetic fibers with an aspect ratio of 105 and integrated sensing capabilities. Through the thermal drawing process, we co-draw conductive nanocomposites based on carbon nanotubes (CNTs) or carbon-loaded polyethylene (CPE) nanocomposites within a thermoplastic elastomer matrix, creating robust, stretchable, and electrically responsive fibers. Magnetic actuators further enhance these fibers, forming a multifunctional system capable of real-time sensing and actuation. We incorporate YOLOv8-based computer vision algorithms to dynamically monitor fiber deformations, including bending, stretching, and compression. The piezoresistive nanocomposites provide a feedback mechanism that enables adaptive responses to mechanical stimuli. These fibers exhibit high sensitivity and mechanical stability under various loading conditions, and rheological analysis confirms the suitability of piezoresistive nanocomposites for thermal drawing. Additionally, fibers can be magnetized in different orientations and woven into fabrics capable of applying forces of up to 7 N using a 5 cm x 5 cm hybrid magnetic textile patch attached to a phantom human hand. These hybrid textiles are programmable for shape morphing, can withstand extreme deformation, and are machine washable. This scalable system presents significant potential for dynamic applications, particularly in the next generation of magnetic medical textiles for rehabilitation and soft orthoses, offering targeted support for joint injuries and controlled force or pressure application.