Single-bout focused-attention meditation (FAM) and open-monitoring meditation (OMM) are assumed to bias metacontrol states towards more persistent versus more flexible processing, respectively. In Experiment 1, we tested whether monitoring and updating of working memory (WM) representations in an N-back task with high (3-back), medium (2-back), and low (1-back) WM demands (varied within participants) is affected by preceding single-bout FAM or OMM meditation (varied between participants and compared with a control group). The results showed that FAM promotes WM performance in the medium (2-back), but not in the high (3-back) or low (1-back) demand condition, whereas OMM did not affect WM performance. A replication of the 2-back condition only (Experiment 2) showed no meditation effect, but a replication of the 3-back condition only (Experiment 3) produced a similar pattern as the 2-back condition in Experiment 1, with FAM promoting performance compared with OMM and the control condition. Taken together, these findings suggest that the single-bout FAM does promote WM performance but only if the capacity demands are neither too high nor too low.