Questions runners ask

Whichever one you already own. Absorb works from Strava, and Garmin, COROS, Suunto, Polar and Apple Watch all sync to it automatically. If your runs appear on Strava, Absorb can read them.

Direct watch connections aren't live yet. Everything Absorb needs is already in what Strava has.

No. Absorb only asks about runs long enough for fuelling to matter. You set that length yourself, and anything shorter is skipped without being counted against you.

Want to log a short one anyway? Open it and do so. A run that went fine takes two taps.

Something on your first logged run, something worth acting on after about five. From two runs in the same condition it will name what has coincided, and it says so in careful language, because two runs is two runs.

Below two runs it says nothing at all rather than guess.

No. Absorb imports 24 weeks the day you connect, then keeps reaching back across a full year of running, working out the terrain, temperature, climb and effort of each run, so the picture of what you've been out in starts full. That part needs nothing from you.

What it can't know is how those runs went. Patterns are built from runs you've marked, so an unmarked run adds to what you've run in but nothing to what you've learned. The learning starts the day you do.

The part you use every week is free, and it isn't a trial: logging, the pattern engine and your full racing record all cost nothing. You pay to plan a race: the course, the pacing, the per-stop fuelling, the GPX write-back and the crew console.

One race is £49, paid once, never expires. Every race is £119 a year. The first 100 members pay £79 a year and keep that price for as long as they stay. See the full breakdown.

No, to all of it. Absorb doesn't sell, rent or license your data, and doesn't use it, or anything derived from it, to train models across users. There's no analytics and no tracking in the product: it doesn't record which pages you open or which features you use.

Three services see a slice each and nothing more. The hosting and database run the service. Open-Meteo gets the time and rough place of a run so the weather can be looked up, never your name or anything you wrote. Anthropic gets the text of race webpages you ask Absorb to read, never your runs, your notes or your health data.

Both, from inside the app, with no email to anyone. Download your data builds a single file with your profile, every run, every note, your race plans and the patterns worked out from them. Anything stored encrypted is decrypted for you. It asks for your password first, because the file is your whole account.

Deleting the account is the same two taps, with a grace period before anything is removed so you can change your mind.

No, and nothing that reads a finished run can. What it can show is which conditions were present when a race went wrong, and whether those conditions had ever appeared in your training.

A race in 30C heat when you'd never trained above 22C is an untested condition. That's a useful thing to know, and it isn't a diagnosis.

Neither. It's a memory and a calibration system. It doesn't count calories, build meal plans, tell you what to eat or write your week. It remembers what you did and what went with it.

It's meant to sit alongside a coach, or your own judgement, as a cross-check rather than an authority.

Absorb doesn't set you a number, and the published range has moved. Guidance clusters around 60 grams an hour for most endurance work, rising to about 90 for long events, and reaching 90 depends on mixing glucose and fructose rather than one source (Jeukendrup 2014 review). Recent trials in trained runners have gone higher: 120 grams an hour, tested over a mountain marathon in elite male runners who had rehearsed the gut for four weeks first (Viribay 2020 trial).

Those are population averages, not advice from Absorb. The number that matters is the one you've actually run on. Absorb never asks you to weigh a portion or keep a food diary: you set your fuelling timing once, and each run records only whether you changed anything and what happened afterwards. Where you want detail, pick products from a library carrying each maker's own published figures and the date they were read.

Absorb is not a medical or dietary service and nothing here is individual nutrition advice.

Absorb can't tell you why. What it can tell you is that race day combines conditions your training rarely combined: longer, hotter, more climb, an earlier start and a higher effort, all at once. Each one may be familiar. The combination may never have happened.

What Absorb does is make that combination visible beforehand rather than afterwards, by showing which of race day's conditions you have runs behind and which you don't.

That you have no marked runs in it, so Absorb has nothing to say about it. A gap in the evidence, not a verdict about you.

Absorb also tracks how long ago you last tested something, so a condition can drift from dialled in, to worth a refresh, to going cold, without anything having gone wrong.

It gives a range, never a single time. Where your training covers the course, the range tightens; where it doesn't, it widens on purpose. A single figure would read as a promise the app can't make.

Once you save a target, your plan does show one time per stop, because that's what you run to; the range is still behind it. Where a race has a cutoff, the chance of not finishing is reported separately, and it will never show you a finish beyond the cutoff.

Yes, and it costs them nothing. You generate a short code and send it. They open it, pick a name and a password, and they're in. No Absorb account, no app, nothing to pay.

They see one race: every stop in course order, where to park, arrival windows, cutoffs, drop bags, the forecast for the hour you're due, and what you need handed over. They see nothing else about you, not your training, your patterns, your other races or your email.

Reading does. Your crew save the whole race to their phone before the start with one deliberate button, and the plan, stops, windows and cutoffs stay readable with no signal at all.

Ticking things off and sending messages need signal, and there's no queue that sends them later, so anything logged in a dead spot has to be redone when signal returns. Your own runner view doesn't have an offline mode yet.

One ultra runner in the UK, and the honesty is enforced in code rather than promised in copy. A runtime guard throws if any response carries a field named rank, score, grade, medal, percentile or a dozen others, so the product cannot ship a readiness verdict even by accident. Where a study's population is known, the change that uses it says so, and where the evidence runs out the app says that instead of filling the gap.

The honest state of it: the pacing constants were fitted against a handful of runners, which is a small sample, and the internal audit says so plainly. That's exactly why the founding members matter. Founding members are the people telling us where it is wrong, on races we have never seen. Your own race history calibrates your own model, and nothing of yours is pooled with anybody else's.

Only if you ask it to. The cycle space is off unless you turn it on, and turning it on is its own consent step naming exactly what will be stored. Every other part of Absorb works without it.

What makes it different is what it refuses to do. Absorb never predicts a phase, never estimates which one you're in, and never adjusts your fuelling or pacing by phase. It records the phases you name, on the dates you give, in your own words. It's encrypted with its own key, never appears in logs, and one tap erases all of it without touching the rest of your account.

Not yet, as of August 2026. Leave an email on the waitlist and you'll get one message on the day it opens, and one about the founding member places. Nothing else.