Morning Grove is a 7–12 minute adaptive AI practice — breathwork, a one-sentence intention, body scan — that tunes itself to your week's actual texture. It is quiet on the days you wake already grounded. It expands, gently, on the days you do not.
A sample 11-minute morning. The exact length and structure adapt to your last 7 days of sleep, your weekly intention, and what you chose to skip recently.
Three slow exhales, a soft chime, no voice yet. The model decides whether to speak this morning at all.
Resonance breathing at 5.5 breaths/min by default, adjusted for resting HRV if a wearable is connected.
A single written sentence, drafted by you, refined by a small on-device model if you ask. Kept private.
Optional, head-to-foot. On stressful weeks the model lengthens this; on calm weeks it disappears entirely.
The ritual is intentionally pruned: most mornings have three segments, not four.
The Morning Grove model reads three signals: your last seven sleep summaries, your last seven self-reported moods (one tap), and a weekly intention you write each Monday. From these it proposes a morning shape. You accept, adjust, or decline any part of it. The model never reads your journal entries — that data lives in Reflect Grove and stays on the device.
Crucially, the model is rewarded for weekly wellbeing — composite subjective wellbeing, sleep quality and sense of agency — not for whether you opened the app today. This single architectural choice changes the kind of mornings it suggests.
Morning Grove is not a streak engine. There is no scoreboard, no monthly leaderboard, no notification on the days you do not show up. We have evidence — in our cohort and in the broader literature — that habit-pressure interventions produce short-term compliance and long-term shame. We have chosen the long-term self-trust path instead.