How people actually practise breathwork: 2026 data from Vayu

Which breathing techniques do people choose?

Box breathing is the most chosen structured technique, with 1,031 sessions from 429 users (10.6% of 9,765 sessions). The 30-second home-screen starter makes up 52.6% (5,132 sessions, 2,355 users). Next come two short Pause routines, Bedtime Calm (597 sessions, 346 users) and Stress Relief (573 sessions, 368 users).

Among the paced patterns, box breathing is followed by 4-7-8 (469 sessions, 245 users), extended exhale (413, 192) and CBT breathing (365, 197). Pranayama is a smaller slice: alternate nostril breathing logged 197 sessions from 101 users and Bhastrika 161 from 116. Cyclic sighing is a recent addition to the app and sits inside "other" because fewer than 20 users had tried it by the cut-off.

By kind of session

When do people breathe?

Mostly first thing in the morning and late in the evening. Among the 298 users whose profile stores a time zone (2,857 sessions), 18.5% of sessions ended between 07:00 and 08:59 local time and 19.5% between 21:00 and midnight, against 2.7% from 01:00 to 04:59. Weekends carried 28.5% of all 9,765 sessions, their expected share.

The single busiest local hour was 08:00 to 08:59, with 11.5% of sessions (328 sessions, 98 users). The late-evening window from 21:00 to midnight covered 557 sessions from 130 users, and 07:00 to 08:59 covered 528 sessions from 114 users. Only about one user in nine has a stored time zone, so treat the local-time figures as indicative.

Methodology and limitations

• Source. The Vayu session log, the table that the iOS and Android apps (and their Apple Watch and Wear OS companions, through the phone) write each completed session to. Read on 2026-09-26 with aggregate SQL only; no row-level data left the database.
• Period. Sessions completed from 2026-01-01 00:00 UTC up to 2026-09-26 00:00 UTC. The first included session is dated 2026-01-04 and the last 2026-09-25. Volume was uneven across the year, with the most sessions in May to July.
• Unit. A session is one saved practice. A user is one signed-in account; the log is keyed to accounts, so the same person on two accounts counts twice and sessions practised without an account are not included.
• Exclusions. Of 9,998 sessions logged in the period, 33 "Self Reflection" journaling sessions (not breathwork), 9 sessions shorter than 5 seconds or longer than 60 minutes (likely accidental taps or timers left running) and 191 sessions from two internal team accounts were removed, leaving 9,765. No biometric field was read: heart rate, HRV and respiratory rate are excluded entirely because this is a behavioural dataset and a known data-quality issue in some older Android records makes those fields unfit to publish.
• Technique names.The app stores names such as "Breathe - Box Breathing" or "Reset - Alternate Nostril". These were mapped to plain technique labels, and renamed variants of the same technique were merged. The 30-second starter is the deep-breathing session on the home screen, which is also the session non-subscribers can try free.
• Suppression. Every published cell describes at least 20distinct users. Techniques below that were merged into "other", and local-time hours below it were grouped into four-hour blocks. There is no country breakdown and no per-user figure of any kind.
• Time zones. Timestamps are stored in UTC and mark when a session ended. Day of week, calendar days and calendar weeks are counted in UTC. Local time uses the time zone stored on 298 user profiles (about 11% of users), who may not be typical of everyone.
• Watch versus phone. The log does not record which device a session ran on, so this page cannot say how much practice happened on the watch.
• Who this describes.Self-selected people who installed a breathwork app and signed in. It is not a clinical sample or a sample of the general population. Vayu's free tier narrowed during 2026, from mostly free to a single free session for non-subscribers, which limits how often people without a subscription could come back. Internal team accounts were excluded; other test accounts may remain because the data does not flag them.
• Not outcomes. These are usage patterns. They say nothing about whether breathwork reduced stress, improved sleep or changed any health measure. For outcomes, see the naturalistic pilot study (N=199, no control group).