MAC/FAC is a model of similarity-based retrieval that addresses three psychological phenomena: (1) structural similarities are weighted more in similarity judgments for items in working memory, (2) superficial similarity is more important in retrieval, and (3) analogical reminders sometimes occur. The model uses a two-stage process: the first stage (MAC) filters candidate long-term memory items using a computationally cheap, non-structural marker, while the second stage (FAC) computes structural matches using the structure-mapping engine (SME). The model is validated through computational experiments showing its ability to model psychological data and outperform an alternative model, ARCS. The model also addresses the challenge of balancing literal and analogical similarity in retrieval, and it is compared to other retrieval models. The FAC stage focuses on structural alignment and inference, while the MAC stage uses content vectors to quickly filter potential matches. The model's design allows for efficient retrieval, with the MAC stage providing a quick, non-structural filter and the FAC stage refining the matches. The model's performance is evaluated through sensitivity analyses and comparisons with other models, demonstrating its effectiveness in capturing both psychological and computational aspects of similarity-based retrieval.MAC/FAC is a model of similarity-based retrieval that addresses three psychological phenomena: (1) structural similarities are weighted more in similarity judgments for items in working memory, (2) superficial similarity is more important in retrieval, and (3) analogical reminders sometimes occur. The model uses a two-stage process: the first stage (MAC) filters candidate long-term memory items using a computationally cheap, non-structural marker, while the second stage (FAC) computes structural matches using the structure-mapping engine (SME). The model is validated through computational experiments showing its ability to model psychological data and outperform an alternative model, ARCS. The model also addresses the challenge of balancing literal and analogical similarity in retrieval, and it is compared to other retrieval models. The FAC stage focuses on structural alignment and inference, while the MAC stage uses content vectors to quickly filter potential matches. The model's design allows for efficient retrieval, with the MAC stage providing a quick, non-structural filter and the FAC stage refining the matches. The model's performance is evaluated through sensitivity analyses and comparisons with other models, demonstrating its effectiveness in capturing both psychological and computational aspects of similarity-based retrieval.