Similar to nonamputees, female athletes with unilateral transtibial amputation (TTA) using running-specific leg prostheses (RSPs) may have worse running economy and higher rates of running-related injury than male athletes. Optimizing RSP configuration for female athletes could improve running economy and minimize biomechanical asymmetry, which has been associated with running-related injury. Nine females with a TTA ran at 2.5 m/s while we measured metabolic rates and ground reaction forces. Subjects used an RSP with a manufacturer-recommended stiffness category, one category less stiff and two categories less stiff than recommended. Use of an RSP two categories less stiff resulted in 3.0% lower net metabolic power (P = 0.04), 7.8% lower affected leg stiffness (P = 6.01 × 10-4), increased contact time asymmetry (P = 0.04), and decreased stance average vertical ground reaction force asymmetry (P = 0.04) compared with a recommended stiffness category RSP. Lower RSP stiffness (kN/m) values were associated with lower net metabolic power (P = 0.02), lower affected leg stiffness (P = 1.36 × 10-4), longer affected leg contact time (P = 1.46 × 10-4), and similar affected leg peak and stance-average vertical ground reaction force compared with higher RSP stiffness values. Subjects then used the RSP stiffness category that elicited the lowest net metabolic power with 100 g, 200 g, and 300 g added distally. We found no significant effects of added mass on net metabolic power, biomechanics, or asymmetry. These results suggest that female runners with a TTA could decrease metabolic power during running while minimizing biomechanical asymmetries, which have been associated with running-related injury, by using an RSP two categories less stiff than manufacturer recommended.NEW & NOTEWORTHY Females with unilateral transtibial amputation can improve running performance through reductions in net metabolic power by using a running-specific prosthesis (RSP) that is less stiff than manufacturer-recommended. Lower RSP stiffness values are associated with greater leg stiffness and contact time asymmetry, and lower stance-average vertical ground reaction force asymmetry. However, we found that adding mass to the RSP did not affect net metabolic power and stance-phase biomechanical asymmetries during running.