Currently there are nearly 2 million people living with limb loss in the United States [1]. Many of these individuals are either transtibial (below knee) or transfemoral (above knee) amputees and require an ankle-foot prosthesis for basic mobility. While there are an abundance of options available for individuals who require an ankle-foot prosthesis, these options fail to mimic an intact ankle when it comes to key evaluation criteria such as range of motion, push-off force, and roll over shape. The roll over shape is created by plotting the center of pressure during a step in a shank-based coordinate system. To address the need for a prosthesis that effectively replaces the ankle's contribution to an able-bodied gait, a biomimetic approach is taken in the design the Compliant & Articulating Prosthetic Ankle (CAPA) foot. The passive CAPA foot consists of four components connected by torsion springs representing the Phalanges, Metatarsal bones, Talus, and Calcaneus. Biomimetic functionality is exhibited by CAPA foot with regards to the roll over shape and a linear relationship between moment exerted and ankle angle, distinguishing the CAPA foot from other ankle-foot prostheses. A mathematical model of the CAPA foot is created to determine the roll over shape a specific CAPA foot geometry would produce and support eventual customization of the 3D printed components. The mathematical model is used to optimize the design to two distinctly different roll over shapes, one with a rocker radius closer to that of the Talus bone and the other closer to the energetically advantageous value of 0.3 times leg length [2, 3]. Compliant and stiff versions of the two CAPA feet were compared to a conventional Solid Articulating Cushioned Heel (SACH) foot and a passive dynamic response foot (Renegade® AT produced by Freedom Innovations). Ten able bodied subjects walked on the Computer Assisted Rehabilitation Environment normally, and then with a transfemoral prosthetic simulator. The study was separated into two experiments. For the second experiment (subjects 6-10), the versions of the CAPA foot had pretension in the dorsiflexion springs. Overall the ankle angles and sagittal plane ground reaction forces of the CAPA foot better mimicked an intact ankle-foot than the existing passive ankle-foot prostheses. Added pretension increased the sagittal plane ground reaction forces and roll over shape radius of curvature and arc length. Nine out of ten participants preferred the CAPA foot and there was a statistical significant difference (F=14.2, p<0.01) between the difficulty level rating given for trials with the CAPA foot versus the existing ankle-foot prostheses. The mathematical model is found to be capable of accurately predicting experimental roll over shape trends and the concept of roll over shape based design is demonstrated. Successful aspects of the CAPA foot can be applied to other ankle-foot prosthesis. The CAPA foot could provide a passive, cheap, and personalizable ankle-foot prosthesis that improves mobility the quality of life for individual’s lacking an intact ankle.