We know only your clock. Imagine what could be noticed with your heartbeat — by something that loves you.
This page uses your local time, your scroll, and nothing else. It will tell you what it knows, every step.
Yokaizen notices a signal, then gives you room to decide what it means.
Your heart is not a metronome
The tiny variations between beats reflect the balance of your autonomic nervous system. When variability drops below your own normal, your body is working harder than usual — often before you'd say "I'm stressed." Drag the slider and watch the variability physically shrink.
Variability is capacity. Yokaizen compares you only to your own 7-day baseline — never to a global threshold.
What the body can and cannot say
Physiology reads arousal — how activated you are — with real fidelity. It is nearly blind to valence: excitement and anger look identical to a heart-rate sensor. Anyone who says their wearable reads your emotions is overselling.
That is why Yokaizen pairs a signal with a conversation. When the body says "activated" and there is no context, she asks an open question instead of asserting a feeling.
You, as a rhythm
The regularity of your sleep timing — independent of duration — predicts rough patches days before they arrive. Drag one night late and watch the regularity score react.
"Rough patch ahead" is a weather forecast, not a diagnosis. Clinical labels never appear in companion speech — by contract, in code.
Breathe with her
When she notices, she doesn't send a paragraph. She offers her rhythm: 4 seconds in, 6 seconds out — six breaths per minute, the resonance frequency of your cardiovascular system. One waveform drives what you see and what you feel. Sixty seconds. That's the whole intervention.
On a phone, your device will pulse with the rhythm where supported.
The same moment, two ways
Stress score elevated. View your trends.
"Hey. Your heart's been working hard for a Sunday night, and you set the quest down halfway. Want sixty seconds with me before sleep?"
Which one would you answer honestly?
Humans don't bond with charts. Yokaizen uses the most powerful bonding loop ever shipped in software — the nurtured creature — and inverts it: this time, the creature takes care of you.
She can be wrong — and should get quieter
PERR — Proactive Emotional Recognition Rate — counts only the moments she names your state unprompted, citing concrete evidence, and you confirm she was right. Spam lowers it. Vague horoscope lines don't count at all. Play the user:
"Your HRV is 22% below your 7-day baseline, it's 11:52pm on a Sunday, and the last quest was abandoned at 40%. Rough edges tonight?"
Risk score below 0.3 → she deliberately stays silent. Friday evenings actively suppress triggers. Silence is a feature.
Receipts, not promises
Tap a card to flip it.
The economics, stated plainly: we are paid when people keep finding it worth paying for, month after month — not per minute of attention, not per ad, not per data point sold. Helping you calmly is the business model; making you anxious would kill it.
We made a memory
The difference between a tool and a relationship is memory. In the app, moments you survive together can become a manga panel of the two of you. Here is yours — from this visit.
The limits
Published verbatim, no animation. Stillness is the design.
- 01
Yokaizen is not a medical device and does not diagnose, treat, cure, or prevent any condition.
- 02
It is not a replacement for therapy or for human relationships — and it will tell you so.
- 03
Consumer wearable readings are trends, not measurements; we compare you only to yourself, and stay quiet when confidence is low.
- 04
The body reveals arousal and energy, not emotions; the companion never claims to read feelings from physiology alone.
- 05
Anticipation is probabilistic; the companion will sometimes be wrong, will say so gracefully, and will learn to be quieter from it.
- 06
If you are in crisis, the companion's job is one thing only: getting you to humans who can help.
The research shelf
Every claim above is load-bearing. Citations back design mechanisms — not clinical claims about Yokaizen.
Task Force of the ESC/NASPE (1996). Heart rate variability: standards of measurement. Circulation 93(5).
Shaffer & Ginsberg (2017). An overview of HRV metrics and norms. Frontiers in Public Health 5:258.
Thayer et al. (2012). HRV and neuroimaging: a meta-analysis. Neurosci. Biobehav. Rev. 36(2).
Kim et al. (2018). Stress and heart rate variability: a meta-analysis. Psychiatry Investigation 15(3).
Radin et al. (2020). Wearable data for real-time illness surveillance. Lancet Digital Health 2(2).
Russell (1980). A circumplex model of affect. JPSP 39(6).
Phillips et al. (2017). Irregular sleep/wake patterns and delayed circadian timing. Scientific Reports 7:3216.
Saeb et al. (2015). Phone sensor correlates of depressive symptom severity. JMIR 17(7).
Lehrer & Gevirtz (2014). HRV biofeedback: how and why does it work? Frontiers in Psychology 5:756.
Zaccaro et al. (2018). Psycho-physiological correlates of slow breathing. Front. Hum. Neurosci. 12:353.
Feldman (2007). Parent–infant synchrony. J. Child Psychol. Psychiatry 48(3-4).
Costa et al. (2016). EmotionCheck: false heart-rate feedback regulates anxiety. UbiComp 2016.
Fitzpatrick, Darcy & Vierhile (2017). Woebot RCT. JMIR Mental Health 4(2).
Maples et al. (2024). Loneliness and suicide mitigation with chatbots. npj Mental Health Research 3:4.
Reis & Shaver (1988). Intimacy as an interpersonal process. Handbook of Personal Relationships.
Cuijpers et al. (2007). Behavioral activation: a meta-analysis. Clinical Psychology Review 27(3).
White & Epston (1990). Narrative Means to Therapeutic Ends. Norton.
Ryan & Deci (2000). Self-determination theory. American Psychologist 55(1).
Forer (1949). The fallacy of personal validation. J. Abnorm. Soc. Psychol. 44(1).
Turkle (2011). Alone Together. Basic Books.
She's waiting to meet you
A companion graded on being right about you — quietly, briefly, and then letting you go.