We examine the impact of the acute manipulation of oxygen availability during discrete phases (active and passive) of a repeated-sprint cycling protocol on performance, physiological, and perceptual responses. On separate days, twelve trained males completed four sets of five 5-s 'all out' cycle sprints (25-s inter-sprint recovery and 5-min interset rest) in four randomized conditions: normobaric hypoxia (inspired oxygen fraction of 12.9%) applied continuously (C-HYP), intermittently during only the sets of sprints (I-HYP) or between-sets recovery periods (I-HYP), or not at all (C-NOR). Peak and mean power output, peripheral oxygen saturation, heart rate, blood lactate concentration, exercise-related sensations, and vastus lateralis muscle oxygenation using near-infrared spectroscopy were assessed. Peak and mean power output was ∼4%-5% lower for C-HYP compared to C-NOR (P ≤ 0.050) and I-HYP (P ≤ 0.027). Peripheral oxygen saturation was lower during C-HYP and I-HYP compared with C-NOR and I-HYP during sets of sprints (∼83-85 vs. ∼95%-97%; P < 0.001), while lower values were obtained for C-HYP and I-HYP than C-NOR and I-HYP during between-sets rest period (∼84-85 vs. ∼96%; P < 0.001). Difficulty in breathing was ∼21% higher for C-HYP than C-NOR (P = 0.050). Ratings of perceived exertion (P = 0.435), limb discomfort (P = 0.416), heart rate (P = 0.605), blood lactate concentration (P = 0.976), and muscle oxygenation-derived variables (P = 0.056 to 0.605) did not differ between conditions. In conclusion, the method of hypoxic exposure application (continuous vs. intermittent) affects mechanical performance, while internal demands remained essentially comparable during repeated cycle sprints.