Sleep mode is built around a single problem: the loop point. A thirty-minute sleep recording set to repeat has a moment where it ends and starts over, and that moment is a change in the room loud enough to matter to a brain that's tuned to notice change rather than volume. Every Official preset below runs the engine that avoids it entirely. StillTune composes the phrase live, on your device, for as long as you leave it running, so there's no seam to hit at three in the morning. These are the team's published Sound DNA recipes, the only sounds on this page indexable by design, each one a different balance of noise floor, weather texture and reverb rather than a different melody. Sleep mode barely uses melody at all once a session settles in for the night.
Most of the difference between these presets comes down to what's layered under the drone. Rainfall Sleep keeps it simple: steady rain with no storm elements, no thunder, no gusts, just consistent patter over the base drone, and it's the most requested type of rain sound on the site for exactly that reason. Distant Thunder and Distant Storm both add weather further out: an occasional low rumble rather than anything close enough to register as change, with Distant Storm running a touch deeper in the reverb tail. If water suits you better than rain, Low Tide and Tide both lean on ocean waves instead: Low Tide pulls the swell back further, brown noise doing more of the work, so it reads like the tide already went out, while Tide keeps a slower, steadier swell under the same underlying drone. Velvet Hush and Hush are the two quietest options, both built on pink rather than brown noise, with Hush trimming the bells back to almost nothing for the lightest touch in the whole mode. And Drift, which loads by default the first time you try Sleep mode, pairs brown noise with a little wind and nothing else competing for attention. Night Wind is its closer cousin: the same texture pairing but leaning further into the wind whenever Drift feels a little too flat.
The engine behind every preset above also changes shape as a session runs rather than holding one texture in place: the phrasing spaces out, the filter gradually closes down toward the low end, and the arrangement sheds material the longer it plays, so what's still on by minute ninety is genuinely thinner than what started the session. That slow narrowing isn't incidental. It's the same auditory-masking principle hospitals rely on when they trial steady background noise against ward or traffic sound, where a consistent low bed raises the level a sudden noise has to reach before it registers as something worth waking up for.
If a preset here still feels a shade too present for how deeply you're trying to sleep, Deep Sleep mode runs the same engine with a heavier drone, longer reverb and bells that arrive once a minute if at all, worth trying before assuming the problem is the music rather than the mode. Nothing on this page downloads as an audio file either. StillTune stores the score and performs it locally, so six hours of sleep music costs a small fraction of what six hours of streamed audio would in data, and Airplane Mode all night works exactly as well as a live connection.