Every focus app now has an AI coach. A sober look at what generative AI can do for your attention - and what is just a chatbot in a calm font. Read →
AI brain-health apps: hype, help, and how to read the scores
Almost every brain-health app now wraps its numbers in AI. Some of that is useful. A lot of it is a confident guess in a calm font. Here is how to tell a score worth trusting from one that is just performing certainty - and what to ask before you hand any of them your data.
The category has exploded for a simple reason: a single daily “brain score” is easy to sell and easy to read. Add a generative-AI layer that narrates the number in plain English and you have a product that feels insightful before it has proven it is. The hard part, and the part that should decide your trust, is what sits underneath the narration.
Why the boom, and why now
Two things collided. Phones and wearables can now passively measure things that used to need a clinic - sleep timing, movement, heart-rate trends. And large language models made it cheap to wrap any of those numbers in fluent, reassuring explanation. The measurement is the slow, careful part. The explanation is the fast, cheap part. The risk is an app that is heavy on the second and thin on the first.
A useful score versus a confident guess
You can usually judge a brain-health score with three questions, and you do not need to understand the maths to ask them.
- Is the input real and measured? A score built on sleep you slept and focus you did is grounded. A score built on a morning mood slider is just your guess, re-priced as data.
- Is the model explained? A good app tells you roughly what moved the number - a short night, a fragmented day. If it cannot say why the score changed, it cannot help you change it.
- Does it admit uncertainty? Sleep stages from a phone are estimates. A trustworthy app says so. A score that reports 73.4 every morning is performing a precision it does not have.
Most of a brain-health score should be carried by things you can measure. A rough sense of a fair weighting looks like this, with the soft, self-reported stuff kept small on purpose.
The data question: where do your numbers go
Brain-health data is about as personal as data gets - how you sleep, how you focus, how steady your attention is over weeks. The honest question is not “is this app secure?” but “does my data leave the phone at all?” Once numbers sync to a server, you are trusting a privacy policy and a roadmap, both of which can change. On-device scoring sidesteps the whole question: there is nothing to breach because there is nothing to send.
A checklist before you trust one
- What does it measure, and is that signal real or self-reported?
- Can it explain why today’s score differs from yesterday’s?
- Does it show estimates as estimates, or hide them behind false precision?
- Does it promise reflection, or imply diagnosis?
- Where do the numbers live - your device, or someone’s server?
| What to check | Green flag | Red flag |
|---|---|---|
| The input | Real, measured signals (sleep, movement, focus time) | A mood slider or a quiz dressed up as data |
| The model | Explains which inputs moved the score and roughly how | A black box that just hands you a number |
| Uncertainty | Admits estimates are estimates, shows ranges or trends | A single confident figure to two decimal places |
| Claims | A reflection to act on, never a diagnosis | Implies it can detect a condition or disease |
| Your data | Computed on-device, nothing uploaded | Account required, numbers sync to a server you cannot see |
Where Soren sits
Soren is built to pass its own checklist. The daily 0-100 Soren Score is computed entirely on your iPhone from three measured pillars - Focus from real deep-work time, Sleep auto-detected from iPhone motion and written to Apple Health, and Physical from HealthKit movement, with the weights published: see how the Soren Score works. The sleep stages are labelled as estimates, because that is what they are. There is no account, no server and no analytics, so the data question answers itself.
84
A Soren Score measured inputs, clear estimate
That last line matters. A score like this is a mirror for the week your attention had, not a verdict on your brain. It is iOS 17+, on the App Store, and it is deliberately narrow: measured inputs, an honest estimate, and nothing that ever leaves the phone.
Weightings and scores shown here are illustrative, not measured fact, and nothing in this article is medical advice or a diagnosis. If we have something wrong, tell us at hello@sorenlabs.xyz and we will fix it.
Soren is a focus and wellness tool - not a medical device, diagnosis, or treatment. The Journal is informational and is not medical advice.
Brain fog feels unmeasurable. Its inputs are not. How you can turn a vague, foggy feeling into one honest daily number. Read →
Attention scores are not a diagnosis. Here is what an ADHD-style score can honestly tell you, what it cannot, and how to use it kindly. Read →