Objectives: We determined whether perceptually-regulated, high-intensity intermittent runs in hypoxia and normoxia induce similar running mechanics adjustments within and between intervals. Design: Within-participants repeated measures. Methods: Nineteen trained runners completed a high-intensity intermittent running protocol (4 × 4-min intervals at a perceived rating exertion of 16 on the 6–20 Borg scale, 3-min passive recoveries) in either hypoxic (FiO2 = 0.15) or normoxic (FiO2 = 0.21) conditions. Running mechanics were collected over 10 consecutive steps, at constant velocity (∼15.0 ± 2.0 km.h−1), at the beginning and the end of each 4-min interval. Repeated measure ANOVA were used to assess within intervals (onset vs. end of each interval), between intervals (interval 1, 2, 3 vs. 4) and FiO2 (0.15 vs. 0.21) main effects and any potential interaction. Results: Participants progressively reduced running velocity from interval 1–4, and more so in hypoxia compared to normoxia for intervals 2, 3 and 4 (P < 0.01). There were no between intervals (across all intervals P > 0.298) and FiO2 (across all intervals P > 0.082) main effects or any significant between intervals × within intervals × FiO2 interactions (all P > 0.098) for any running mechanics variables. Irrespective of interval number or FiO2, peak loading rate (+10.6 ± 7.7%; P < 0.001) and duration of push-off phase (+2.0 ± 3.1%; P = 0.001) increased from the onset to the end of 4-min intervals, whereas peak push-off force decreased (−4.0 ± 4.0%; P < 0.001). Conclusions: When carrying out perceptually-regulated interval treadmill runs, runners adjust to progressively slower velocities in hypoxia compared to normoxia. However, only subtle constant-velocity modifications of their mechanical behaviour occurred within each set, independently of FiO2 or interval number.