Primary drying is the bottleneck of pharmaceutical lyophilization, as it represents the most time and energy-consuming stage and largely determines the quality of the final product. The common practice runs the process under pre-calculated input profiles while considering average drying dynamics for the batch. This results in conservative and underperforming cycles. In this regard, the paper proposes and experimentally validates an in-line control strategy for primary drying. It explicitly accounts for intra-lot drying heterogeneity using three models, each one calibrated for a reference vial position in the chamber. Two separate experiments evaluate the proposed framework: offline and in-line process control. Validation on a pilot-scale unit shows a cycle time reduction of the proposed approaches compared to standard drying protocol at constant operating conditions. The input profiles obtained by offline control lead to a 16% shorter cycle time; however, the lack of feedback during the cycle impedes addressing potential parametric variations and model mismatches. Conversely, in-line control yields a reduction of 13% while allowing careful monitoring of the product status, ensuring a more dependable cycle. In both cases, the visual quality inspection confirms the acceptability of the final product without visible degradation.