EnduCloud

04/09/2026

One of you is abroad on holiday. One works shifts and can only run at 22:00. One moved to Berlin two years ago. You still want to run together.

Together is now rolling out in the EnduCloud pilot: one event, any number of places. Everyone runs their own route, either at the same moment or whenever it suits them inside a window the organizer sets, and a single live board shows the whole group moving through it.

Three things it is for:
A group of friends, scattered. Four people, four cities, one Sunday long run. You watch each other's progress as it happens, and you all finish the same event.
Organized community events. An event like a national memorial run where thousands take part without a single start line. Each participant runs their own race, at their own hour or all at once, and one public board holds the whole thing together. Sharing a link is the entire registration flow.

Live tracking at a real race. Your family follows you through the course from wherever they are, in a browser, with no account and nothing to install.
Your run counts however you recorded it: watch, phone, file import, health platform. The public board shows progress only, never a map trail or a physiological data, and each participant chooses whether their name appears at all.
Rolling out now on the web and Android and iOS following soon. Early, and we would rather say so.

02/09/2026

We went looking through our own database for a signal we suspected was there, and found something worth sharing, including a part that almost fooled us.
The question: when an athlete is getting sick, digging a hole, or simply not absorbing their training, is there a signal in their training data, before they think to mention it?
We have two numbers for the same session. What the athlete said it felt like (1–10), and what our load model (STRSS™) measured it cost, fitted per athlete against their own critical power and threshold, not a generic formula. Over twelve months, 2,322 sessions carried both, from 73 athletes with enough of each to compare.
The first result was a disappointment that turned useful. Perceived effort is a weak predictor of measured load: across the entire 1–10 scale, load moves only about 1.6×. Duration explains roughly five times more. The reason is session structure, as the median session in every rating bucket was about an hour, so a session rated 9 isn't an hour at 9, it's an hour with hard intervals in it. Effort tracks the peak; load accumulates over the whole.
Which means the difference between the two carries information neither carries alone. When that gap stays elevated across consecutive sessions, we checked it against two measurements that had no part in producing it: resting heart rate ran +0.19 sd, HRV −0.32 sd. Two independent autonomic markers moving the way the hypothesis predicts.
The catch is the most useful thing here. Include ratings below 3 and the resting-HR association flips the wrong way - not noise, arithmetic. On a coarse integer scale a step from 1 to 2 doubles any ratio built on it. A metric derived from a low-resolution input has a floor below which it misleads rather than merely wobbles. If you're building on a 1–10 human rating, find that floor before you trust the output.
Effect sizes of 0.2–0.3 sd from 30 episodes are a nudge, not a diagnosis. So it's rolling out in our pilot as an advisory – supportive input next to HRV, resting HR and sleep. It proposes; it never silently changes a plan. Every flag is stored with its numbers so we can score the calls against what actually happened.
Two of our three findings were negative results. That's exactly why the third is worth anything.

22/08/2026

Working on EnduCloud puts me in conversation with a wide range of people: athletes, coaches, physiotherapists, physicians, sport scientists. What has started to stand out is who is never in those conversations.

Two groups are silent. Healthy sedentary adults, who aren't in the room because lack of exercise and its impact on their body is not yet a problem they feel. And less-healthy sedentary adults, who are in the room only by proxy, represented by a clinician describing them through a professional lens.

These are exactly the people with the most to gain from simply becoming more active. The first group is busy with the rest of life, sheltered under an umbrella of not knowing what two or three decades of inactivity are doing on the inside, until the symptoms arrive in their 40s and 50s. The second is stuck because the system holds itself to a standard of care it cannot deliver at scale - time, cost, expertise, level of supervision - and when it cannot meet that standard, it does nothing. Nothing is the most expensive option on the table.

What we are building is the middle: individualised, progressive guidance that starts where the person actually is, adapts to what their body reports back day by day, and gives the professional around them visibility without demanding an hour per patient per week. Not a replacement for a clinician or a coach. A way for one of them to carry a caseload instead of a handful.

In practice that means plans that begin at walking and need nothing more than shoes and a phone. Targets expressed in something the person can feel, rather than a number they need a lab to interpret. And a system that reads sleep, resting heart rate and reported symptoms and adjusts, so that "safely" is a property of the plan rather than a warning printed underneath it.

We are still early. This is a pilot with real users, not a finished product. But the gap here is not a technology gap. It is a supervision-capacity gap, and that one is solvable.

20/08/2026

Last week we argued that data alone is a dashboard, and a loop is a coach. Here's the part we left out: a loop is only as good as what it can see.

You train for about an hour a day. Your body adapts during the other twenty-three, while you sleep, recover, stress, and live. A coach who only sees the workout is coaching with one eye closed.

That's why EnduCloud connects the devices people actually wear: Garmin, Polar, Suunto, Wahoo, Google Health, and listens both during sessions and around the clock: sleep, HRV, resting heart rate, daily load and recovery. Wear more than one? Using wellness Rings and bands? Also fine. The signals are reconciled into a single profile, so a watch on your wrist, bike computer on your bar, and wellness band when you don’t train, don't become multipart-athletes.

That around-the-clock profile is what our AI coach prescribes from. Tomorrow's session isn't picked from a template. It's adjusted to how you actually slept, recovered, and absorbed the last one.

And here's where we're heading, openly labeled as roadmap: an optional layer of more advanced monitors like continuous glucose, blood pressure, dedicated sleep tracking, for people who train for health, and for training that happens under professional supervision, always alongside a medical team. EnduCloud is not a medical device and doesn't diagnose; it turns signals into context, and context into a better plan.

The loop stays the same. We're widening what it sees.

15/08/2026

I swear this is not an invited article

The European Heart Journal published a review on wearable devices and cardiovascular health (Hughes et al., Eur Heart J 2026;47(18):2130-45, open access), and it reads like the design meetings we have every week.
What it says, condensed:
• Wearables on their own produce modest, short-lived behaviour change. Paired with behavioural coaching and professional consultation, the effect holds.
• Detection is not benefit. Smartwatch AF alerts reach impressive predictive values, yet population screening has not shown a stroke reduction. An alert is a starting point, not a diagnosis.
• HRV and cuffless blood pressure are not ready to drive clinical decisions on their own - reference ranges are missing, and a step counted by one vendor is not the same step counted by another.
• Their proposed path forward: human-in-the-loop models, where coaches and allied health professionals review trends before anything escalates to a physician.

That last point is precisely the layer EnduCloud was built to be, long time ago, before anyone was speaking about it - the training and behaviour layer that turns continuous data into coaching, alongside the medical team, not instead of it. We are deliberately not a diagnostics product, and this review is a good articulation of why that line matters.

What makes the paper useful is its honesty about the limits. Worth your time if you work anywhere near wearables, prevention, or endurance health.

https://academic.oup.com/eurheartj/article/47/18/2130/8540180

Abstract. Wearable devices are transforming cardiovascular medicine by enabling continuous monitoring of physiologic and behavioural measures outside of tr

15/08/2026

In June 2025, the NEJM published the CHALLENGE trial: colon cancer survivors assigned to a structured, supported exercise program had meaningfully better disease-free survival than those given education alone. Not a supplement. Not a device. A training plan, with someone making sure it actually happened.
The evidence in exercise oncology has been building for two decades. Regular physical activity is associated with lower risk of several cancers, better tolerance of treatment - less fatigue, preserved muscle and fitness - and lower recurrence risk in the most-studied cancers. The ACSM's guidance for people in and after treatment is concrete: roughly 150 minutes of moderate aerobic work per week, plus resistance training.
The gap has never been the evidence. It's adherence. A person mid-treatment doesn't need a generic plan - they need one that adjusts to how their body is actually doing this week, and doesn't punish them for a bad one.
That's the problem EnduCloud is built around: personalized training that listens to your body's signals, including sleep, HRV, readiness, how the last session went, and adapts the next one accordingly. Structure when you can, grace when you can't. It's what keeps people moving month after month, and consistency is where all of the benefit lives.
We're also developing supervision features with healthcare professionals so that, in time, structured activity for clinical populations can happen under proper clinical oversight. That work moves at the pace of clinical sign-off, as it should.
If you're in treatment or recovery: talk to your medical team first, always. And then - move. It's one of the best-evidenced things you can do for yourself.

10/08/2026

Here's a number worth sitting with: in most high-income countries, the last ~10 years of life are spent in poor health. 🧡

We got very good at adding years. Less at making them good ones. And that part isn't in a supplement bottle. The evidence points at things you can train: aerobic fitness, strength, recovery, and - the most underrated one - doing it with other people.

That's exactly what EnduCloud is built for, end to end:
📊 It knows how you are - sleep, recovery and training load from whatever watch you already wear
📋 It knows what to do - plans that mix aerobic, strength and mobility, and adapt when your body needs easier days
🏃 Guided live workouts - everyone follows on their own device, at their own pace, together
🧑‍🏫 A coach stays in control - seeing the whole group, adjusting each person individually 👓 Comfort View - a simplified mode for anyone who wants big, clear guidance

More life in the years. That's the goal.

We're onboarding pilot communities now - if you lead a training group (or want to start one), reach out.

06/08/2026

Exercise is medicine 💊 - but medicine works because it has a dosage, a record, and follow-up. Training advice usually has none of the three.

New in the EnduCloud pilot this week:
🎯 Pace sessions with exact targets and an honest tolerance (±3 s/km) - in units your own watch can actually measure
📊 Every session scored against the prescription, so you know how well you executed, not just that you showed up
🤒 Feeling something coming on? If your morning resting HR climbs above your personal baseline, your plan adapts the same morning - rest, swap, or an easier session, with the reasoning shown

One AI coach, watching how you are and adjusting what you do. Open the app and check today's session 🏃

(Rolling out now in the pilot - tell us what you think.)

01/08/2026

Exercise is one of the best-evidenced interventions in medicine, and one of the worst-delivered.

What the research actually shows, from the earliest warning signs through to active disease:

• Before any diagnosis – low cardiorespiratory fitness carries mortality risk comparable to or greater than coronary artery disease, diabetes or smoking. Between the least and most fit groups: an 80% difference in mortality risk, with no ceiling to the benefit (JAMA Netw Open 2018, 122,007 people).

• Heart disease – exercise-based cardiac rehab cut all-cause hospitalisations by 42% and heart attacks by 28% within 6–12 months. The fair caveat: cardiovascular mortality only improves beyond three years, not in the first (Cochrane 2021).

• Cancer – the minimum useful dose is now 90 minutes a week. Training during and after treatment is broadly safe and improves fatigue, physical function and quality of life (ACSM 2019).

• Metabolic – 58% lower progression from prediabetes to type 2 diabetes through lifestyle change, versus 31% for metformin (Diabetes Prevention Program).
Here is the part that gets lost: every one of those numbers is a population average. None of them tells you what this person should do on this Tuesday, on a beta blocker, three days after a treatment cycle, or after a bad night's sleep.

The dose has been settled for years. Making it personal, and keeping track between clinic visits, is the part still waiting to be solved. That is the direction we are taking EnduCloud.

The clinical layer is in pilot now. We are not enrolling patients at scale, but clinicians who want to try it on themselves, or with a small sample group, are welcome to get in touch: [email protected]

28/07/2026

Following input from our coaches, we have added live coaching to EnduCloud.
Your coach can now see your session while it's happening, and talk to you during it.
What the coach sees:
• Your heart rate as your own zone, not a number they have to decode
• Pace, grade-adjusted pace, elapsed and distance
• Where you are, on a map, with your trail behind you
• A nudge if you go quiet, stop while the group keeps moving, drop off the back, or your battery is about to die
What you get:
• A note from your coach mid-session, on screen and spoken aloud
• One tap for the common ones - ease off, pick it up, 5 min to go, take a drink, good work, next rep
• Or their actual voice, in a clip under ten seconds
• Notes wait their turn, so they never talk over your interval countdown
And you keep your Garmin. It records and syncs exactly as it does now. Turn on Broadcast Heart Rate, put your phone in your pocket, and that's it.
Your call, always: sharing is off by default, your heart rate and your location are separate choices, you can switch it off mid-session in one tap, and you can always see when your coach is watching.
One thing we want to be straight about - this is your coach seeing you, not a safety system. It needs your phone to have battery and signal.
Now rolling out. Open the app and have a look.