Help
For the full walkthrough - every setup step in order, with a short video each - see the setup guide. The questions below cover what people ask most often.
An iPhone 11 or newer running iOS 26 or later, and an Apple Watch Series 4 or newer running watchOS 10 or later. The Watch is essential - it is where tremor and dyskinesia are sensed. You do not need a recent one: a Series 5 or a first-generation SE works fine.
No - this is expected. Tremor sensing needs a few days of regular Watch wear before patterns appear. Wear your Watch through your normal days, and Kampa will start showing data as it accumulates.
If you still see nothing after 3-4 days of wear, the usual cause is a permission rather than a fault. Check that Kampa is installed on your Apple Watch, and that Motion & Fitness is on in Settings → Privacy & Security → Motion & Fitness. Without it the Watch returns no tremor readings at all, and the app cannot tell that apart from a calm day - which is why it now says so directly instead of showing you an empty chart.
No, but open it every few days. Your Watch stores about a week of tremor data and hands it to your iPhone when the app opens. If the app stays closed longer than that, the oldest days can be lost. Kampa does not send notifications yet, so nothing will prompt you - opening it a couple of times a week is enough. A reminder when your Watch has not been heard from in a while is the first notification we plan to add, and we are working toward fully automatic background sync.
Two upfront: Health access (to read sleep, exercise, heart data, medications, and to save what you log), and Motion access on the Watch (tremor and dyskinesia sensing). Three more are asked for only at the moment you first use the feature - the microphone and speech recognition for voice logging, and the camera for scanning a food barcode. Kampa does not ask for notification access, because it does not send notifications. Nothing is requested that is not used.
Either way works, and you do not have to choose. Tap + → Medication in Kampa: name each substance once, and after that a dose is a single tap from a list ordered by what you are most likely taking now, with the time you last took each one on the row. Kampa also keeps reading anything you log in Apple Health, so doses recorded there still appear on your timeline.
If you already had a medication history in Apple Health, Kampa imports it once, so nothing you have recorded is left behind. Doses logged in both places are matched, not counted twice.
Kampa never pre-filters to "Parkinson's" medications. Anything you take - a supplement, a blood-pressure tablet - can be logged and correlated, because the point is to find out what actually affects you.
It is a None-to-Strong severity estimate produced automatically from your Apple Watch. The labels echo the severity language clinicians use, but the number is a passive estimate from wrist sensing - not a doctor's exam, and not a clinical score. Use it to compare your own days against each other, not to compare yourself against anyone else.
Each point on the line is your typical tremor level over a short window - about ten minutes - so the chart shows how your tremor actually rises and falls through the day rather than one flattened number. When you touch the line, the callout reads that same window, so the number and the curve always agree. The window is kept short on purpose: tremor can change quickly, and a longer average would smooth a real ten-minute surge down below what you felt.
The figure beside "Tremor" at the top of Review is a plain average across the entire day, calm hours included. Tremor is often either near zero or clearly elevated, with little time in between - so a whole-day average settles at a level you rarely actually sit at, below your real peaks. Read the chart for the shape of your day; treat the daily average only as a rough way to compare one day against the next.
Different apps measure different things. Kampa's chart shows tremor intensity - how strong the tremor was, through the day. Some other apps report duration - how many minutes of tremor occurred. Both are legitimate lenses; a day can have long mild tremor or brief strong tremor. Neither number is wrong; they answer different questions.
Dyskinesia refers to involuntary movements that can occur as a side effect of levodopa medication, usually at peak dose. Not everyone experiences it - if you don't, a flat line is the correct reading, not a broken feature. Wrist sensing can also register some ordinary movement as possible dyskinesia, so Kampa is deliberately conservative about what it shows.
Because pharmacologically they often are. Tremor tends to peak before a dose (medication worn off) and dyskinesia at peak dose. Seeing them rise and fall in opposite rhythm around your doses is the levodopa cycle made visible - many people find this one chart explains their day better than anything else.
On today's Review, Kampa projects the likely shape of your day - your probable on and wearing-off windows - from the doses you've logged and how each of your medications has tended to act for you, learned from your own history rather than a textbook drug curve. It is a heads-up for planning, not a dosing instruction. As always, decisions about medication are for you and your neurologist.
The panel has two modes and always tells you which one you are looking at. While a dose is still inside its expected window, you get on and off windows. When no dose is active - a day you have not dosed yet, or hours after the last one has cleared - it stops projecting windows and shows your own typical range instead, and says so. Stretches are then marked usual or better or worse than usual, measured against your own baseline rather than anyone else's.
Stretches where your Watch was off your wrist are drawn as no watch data rather than quietly filled in.
Because it only claims what it can explain. A projected window is tied to a specific dose whose behavior Kampa has measured enough times to model. Outside those windows it falls back to your typical range. Kampa deliberately does not stretch a dose's influence past the point where the evidence runs out, which is why the colored windows sometimes cover less of the day than you might expect.
Observations describe today: "Tremor eased 68% in the hour after your 8 AM dose." They are honest descriptions of what co-occurred, not proof of cause. Insights are patterns that hold up across many days and pass statistical checks before you ever see them. Observations are the daily read; Insights are the accumulated evidence.
Emerging, moderate, and strong reflect how much evidence supports a pattern - how many days, how large the effect, and how consistently it repeats. Kampa shows the basis plainly, so a medication card will say something like "measured on 160 of your own 252 doses" and you can see how much data sits behind the claim. If a pattern stops holding, its card is downgraded or retired. We would rather show you nothing than something that is not real.
Because that is a result too. If Kampa checks whether caffeine moves your tremor and the answer is "the difference sits inside your usual daily variation," it says exactly that rather than dressing a non-finding up as a discovery. Most consumer health apps only show you the hits. A pattern engine that never reports a miss is not measuring anything - it is just narrating.
Because Kampa applies real statistical standards before claiming anything, and honest standards mean fewer, better cards. Some things genuinely have no effect on your symptoms - learning that is valuable too. Cards appear when evidence clears the bar and retire when it no longer does; the screen reflects the current state of evidence, never a highlight reel.
Because they behave differently, and averaging them together hides both. If you take Sinemet and Mucuna Pruriens, Kampa measures each one separately - how long it holds, and how many of your own doses that estimate rests on - rather than reporting a single blended number that describes neither.
Whether a card exists depends only on whether you take the substance. Kampa does not decide in advance which treatments deserve measuring, so traditional and unproven treatments are measured exactly like anything else. That is not an endorsement of them; it is the opposite of an opinion. If you take it and there is enough data, you get the number, and what you conclude from it is yours and your neurologist's to decide.
It compares how long your doses tend to hold against how long you are awake. If each dose covers roughly three hours and your doses do not span your waking day, the uncovered stretches add up, and that total is what the card reports.
Uncovered does not mean OFF. It means no dose was inside its expected window - not that you were symptomatic. You may well feel fine during an uncovered stretch. The card is measuring your dosing schedule against your day, not measuring how you felt.
A single number for how you slept, built from your own history rather than a population average, so it answers "how did I sleep compared with my normal?" It can differ from other sleep scores you have seen - Apple Health does not publish one, so this is Kampa's own calculation from the sleep stages your Watch recorded, and other apps use their own formulas.
Yes. The summary tiles at the top of Review are doors - tap sleep or HRV and it opens into that measure's own history, with today marked against your usual range so you can tell an unusual day from an ordinary one. Scrubbing the chart moves every panel on the screen together, so you can line one signal up against the rest.
Yes. Apple Health pools data from every app and device that writes to it, which sometimes means duplicates or a source you do not trust. Settings lets you exclude a data source app-wide, and Kampa will ignore everything from it.
No. Never. Kampa may surface an observation worth discussing - for example, that your afternoon doses seem to act more slowly than your morning ones - and help you prepare a summary for your neurologist. What to do about it is a decision for you and your doctor.
On your iPhone and in your own private iCloud. Kampa has no servers. We cannot see your data - not because of a policy, but because of how the app is built. Nothing is sold, shared, or used for research without your explicit consent.
No. Everything Kampa does today runs on your device. The patterns you see are computed on your iPhone, and no health data is sent anywhere to be analyzed - there is no AI feature in the app to opt into yet.
If we add one - plain-language explanations of your own patterns are the obvious candidate - it will be opt-in, and it will send only derived summary statistics: never your raw health data, never your identity. It is written here so you can hold us to it.
Yes, always. Kampa exports your data as CSV files you can keep, inspect, or share. It is your data; export is free and always will be.
It stays in your iCloud and your Apple Health store, under your control. Deleting the app removes Kampa's copies; your Health data remains yours in Apple Health.
Kampa will tell you. Because your private iCloud is the only place your record is backed up, a full account is the one failure that could quietly cost you months of data - when iCloud fills, backing up simply stops, with no warning from anywhere else. Kampa checks for it and shows a banner on your Review screen reading "iCloud storage is full - new data has stopped backing up."
Freeing up space, or moving to a larger iCloud plan, starts it again on its own. Tremor sensing and everything on your phone keep working throughout - it is the backup that pauses, not the app.
For the complete details, see the full privacy policy.
The simplest way is inside the app: tap + on the Review screen, then speak. Say "5 almonds with tea," "bloated after lunch," or "ten minutes of breathing" and Kampa sorts each into the right place - food, a symptom, or a mindfulness session. It reads back what it understood so you can confirm before anything is saved. Siri can also start food and mindfulness logs hands-free ("log a 20 minute meditation") on the occasions it picks the request up, though the in-app microphone is the reliable route. It was designed for days when typing is hard.
No. Everything can be typed, and packaged food can be scanned. Voice exists because typing is hard on some days, not because it is the only way in.
Yes. Tap the speaker icon beside Observations on any day and Kampa speaks that day aloud: how you slept against your own recent nights, your average tremor with the times it peaked and eased, what each dose did, then everything else it noticed. Tap again to stop. It never speaks on its own - there is no timer and no notification behind it.
Worth doing once: install a better voice. iOS ships a basic one that sounds flat. In Settings → Accessibility → Spoken Content → Voices, download an Enhanced or Premium voice - Kampa picks up the best one you have installed automatically, with nothing to configure on our side.
In Log Food, tap Scan a package barcode and point the camera at the barcode. Kampa fills in the product name and its nutrition, and records the attributes it can correlate against - protein, sugar, fat, fiber, caffeine.
It searches 439,082 products in an 11.8 MB file bundled inside the app, so there is no network call and nothing about what you eat leaves your phone. It is fastest for anything that came in a wrapper.
Type or say it instead, and Kampa classifies it the same way it classifies any other described meal. The bundled catalog covers packaged groceries well, and it is thinner on regional and imported foods - South Asian products especially. Home-cooked food has no barcode at all, so describing it is the normal path, not a fallback.
Six things it can correlate against: protein, starch, fat, fiber, sugar and caffeine. Protein and starch matter most - protein competes with levodopa for absorption, and starch is the other side of that same mechanism. Kampa does not count calories or score your diet; it only records the handful of components that could plausibly change how a dose lands.
Where the nutrition data behind a food does not record enough to work a value out, Kampa marks it unknown rather than assuming there is none of it.
No - and it is deliberately different from "not recognized." No tracked nutrients means Kampa found the food and it carries none of the six things above. Plain rice is the clearest example: it is in the database, and what it mainly contains is carbohydrate, which Kampa does not track as its own signal because it is on almost everything and so cannot separate one meal from another.
Not recognized means the word did not match anything. Try a plainer or more common name for the same food - and if it is a dish that should be there, tell us and it gets added.
Yes. In Settings → Medications you will see everything Kampa can see - what you log in the app and what it reads from Apple Health - and a switch takes any of them out of the analysis. The dose stays on your timeline and in your record; it simply stops being treated as something that might explain your tremor.
This matters more than it sounds. Kampa sets aside anything you did close to a levodopa dose before it looks for a pattern, so a medication wrongly counted as Parkinson's medication quietly removes data it should have kept.
Tap +, then Symptom. There are seventeen, covering the non-motor side of Parkinson's that a tremor reading never sees - constipation, fatigue, low mood, urinary symptoms, dizziness, body ache, memory lapses, loss of smell, sweating, drooling, swallowing, speech and others. You can also just say it - "bloated after lunch," "nausea this morning."
Each one takes None, Mild, Moderate or Severe. The first of those four matters more than it looks: without a way to record a good day, a symptom's history is made only of bad ones, and the app becomes structurally blind to the thing you most want to see - a symptom stopping.
Log when something is worth recording. A quiet day needs no entry. There is no daily checklist to keep up with and nothing to fall behind on, which matters more than it sounds: a tracker you feel guilty about is one you stop opening.
Where each one is stored depends on Apple. Eight have a matching Apple Health type and are written there, so they appear in Health alongside everything else. The other nine have no Health equivalent, so Kampa keeps them itself - on your device and in your own private iCloud, exactly like your tremor record. Nothing about how you log them changes, and both kinds appear on the same timeline and feed the same insights.
Constipation and gut motility are among the most common non-motor patterns in Parkinson's, and gut motility also affects how well levodopa is absorbed - which is why these sit on the same timeline as your doses rather than in a separate diary.
Not yet - voice covers food, mindfulness and symptoms today. A dose is two taps in the app, so it was the least urgent of the four, and it is on the list. Apple's own rules used to make this impossible; that is no longer the reason.
Yes. Kampa keeps a list of the therapies you actually use - you name each one once, then log sessions against it with a start and end time. Sessions land on your timeline beside your doses, so a treatment sits in the same picture as everything else that day.
Because the name lives on the therapy rather than on each session, renaming one updates every session you have already logged. If you stop something, archive it: it disappears from the list you pick from, and every session you logged against it stays exactly as it was. Nothing you have recorded is ever rewritten by a change you make later.
Six short tests you run on the phone, most taking ten seconds per hand. They are the deliberate counterpart to everything else Kampa does: the watch measures your tremor passively, all day, without you thinking about it. These you choose to do, at a moment you pick.
Each one is matched to a movement a neurologist checks for by hand, so a result means the same thing your clinic visit is looking at: finger tapping, hand rotation, leg agility, postural tremor, rest tremor and a drawn spiral.
Two reasons. The first is that several of them measure a different thing entirely: bradykinesia - slowness of movement, and the way movements get smaller as they repeat. It is a separate motor sign, it does not track tremor, and passive sensing does not capture it. Someone can have a quiet tremor day and a slow one at the same time.
The second is that a watch is worn on one wrist and gives back a single number, so it cannot tell tremor at rest from the kind that only appears when you hold a position against gravity. Those are different signs with different meanings. The phone checks can tell them apart, because you choose the posture.
Tapping. Set the phone flat on a table. Tap two targets with your index finger, one hand at a time, for ten seconds each.
Rotation. Sit down and get comfortable. Hold the phone flat against your open palm with your arm out in front, then turn it palm up and palm down, as fast and as fully as you can, for ten seconds each hand. Only your forearm moves. Do it over your lap or a sofa, in case the phone slips.
Leg agility. Sit with your feet flat. Hold the phone against your thigh and stomp your heel on the floor, lifting the leg fully each time, as fast and as high as you can for ten seconds a leg. Go all out every time - a slower run is not comparable to a faster one.
Postural tremor. Hold the phone out in front of you with your arm extended and stay as still as you can. Ten seconds a hand.
Rest tremor. Rest the phone in your lap or on the arm of a chair, hand relaxed and supported, and let it be still. Ten seconds a hand. The point is that you are not holding anything up.
Spiral. Trace a spiral on the screen with your finger, from the centre outwards. This one is not timed - draw it at whatever pace feels natural.
All the hand checks ask for both hands, in one sitting. Parkinson's is usually asymmetric, so the gap between your two hands is often the most informative thing on the screen - and it only exists if both were measured at the same moment.
One beep means start. Two beeps mean stop. Every timed check uses the same pair, so a session sounds the same whichever one you open.
They exist because of Rotation: the phone is on your palm at arm's length, so you cannot see the countdown, and a vibration traveling down an outstretched arm to a palm that is itself turning over is easy to miss. Sound is the one channel that survives the posture. It is the count that carries the meaning rather than the pitch - two events are unmistakable against one, which is not true of a high tone against a low one.
Tapping reports how many taps you made, your accuracy, and how spread out they were. Rotation reports how many turns you completed and your amplitude - how far your hand turned on each flip, on average. Leg agility reports how many stomps you managed and how high you lifted. Postural and rest tremor report how much your hand moved and how fast it shook. Spiral reports how far your line strayed from a smooth one.
Each number is shown against your own usual range, not against anyone else's, and both hands appear on the same chart so you can read the difference between them directly. The screen also notes which medication you last took and when, so a result reads as "compared with my usual, at this point after a dose" rather than as a score on its own.
Kampa picks for you, and it asks for the check and the moment your record has least of - not the one that would be easiest to compare. Filling in the gaps is what eventually lets it answer whether your medication is doing anything, and asking only for comparable moments would pile every reading into one part of your dosing day, where that question can never be answered.
There is no schedule and nothing to fall behind on. A single check tells you very little, because the whole comparison is against your own usual range - so a handful spread over time is worth far more than one taken carefully. Since each result records where you were in your dosing day, checks taken at a similar point relative to a dose are the ones that compare cleanly.
These are a measurement of your own movement over time. They are not a diagnostic test, not a score against other people, and not something to pass or fail.
No. Kampa works fully without one. If you choose to wear an over-the-counter continuous glucose sensor (such as Abbott Lingo or Dexcom Stelo - no prescription needed), Kampa can display your glucose curve time-aligned under your tremor chart, with doses and meals marked across both.
Levodopa is absorbed and reaches the brain through transport systems that food - especially protein and large meals - can compete with. Your glucose curve is an objective record of your body's meal response, which can help make sense of days when a dose seemed slow or weak. Kampa treats this as an area to observe honestly, not a promise.
Enable Apple Health sharing inside your sensor's own app (for example Lingo). Kampa reads glucose from Apple Health automatically - no separate pairing.
Nothing right now - Kampa is free to use. Sensing, charts, daily observations, logging, and data export are all included.
Kampa is built by Bhav Bhasin, a product builder who has spent 25+ years on large-scale technology and now builds independently, working closely with the Parkinson's community. Kampa began as an attempt to answer questions that existing tools would not: not "what happened," but "what actually affects me?"
Kampa (कम्प) is Sanskrit for "tremor," from Kampavata - the term ancient Ayurvedic medicine used for what modern medicine calls Parkinson's disease. We chose a name that faces the condition honestly.
Parts of it. Kampa's lifestyle logging and correlation tools apply broadly, but the Watch's tremor sensing is tuned by Apple for Parkinson's resting tremor, and some features (like dose-response analysis) are specific to Parkinson's medication. We have not yet validated Kampa for essential tremor, and we won't claim it works until we have.
No. Kampa is a personal wellness and symptom-tracking tool. It is not a medical device and does not provide medical advice, diagnosis, or treatment. Always consult a qualified healthcare professional about any medical condition.
support@kampa.health - it lands with the person who builds the app, usually within a day or two. Ideas from people living with the condition have shaped most of what Kampa is.
Email support@kampa.health - it reaches the person who builds Kampa, usually within a day or two. The support page has a few quick checks worth trying first.