You have four years of data about yourself. What did it ever tell you?
Your watch counted every night. Your lifting app kept every set. Your macro app has ten thousand meals. Each one holds a column of you, and none of them holds the two rows that explain the rest: what you took, and how you felt.
Vayl reads them all in and puts everything on one axis, so the answers you never got are the ones you can finally read.
The web terminal is in closed testing
You never had a capture problem.
You have been measuring for years, across four subscriptions. What you never had was the place where all of it lands together, including the two rows nothing on your wrist can record.
- Heart & recoveryheld VaylheldWristheldRingnot heldLifting appnot heldMacro app
- Sleep architectureheld VaylheldWristheldRingnot heldLifting appnot heldMacro app
- Sets, load & volumeheld Vaylnot heldWristnot heldRingheldLifting appnot heldMacro app
- Macros & intakeheld Vaylnot heldWristnot heldRingnot heldLifting appheldMacro app
- What you tookheld Vaylnot heldWristnot heldRingnot heldLifting appnot heldMacro app
- How you feltheld Vaylnot heldWristnot heldRingnot heldLifting appnot heldMacro app
Vayl reads your wearables through Apple Health: route, and per-sample heart rate, power and cadence. Not a summary.
None of this gets typed at a desk.
It gets logged between sets, in the kitchen, on the road, with one hand and no signal. The phone is where the record happens. The desktop is where it finally means something.
Strength
44 sets · 7 exercises
Cardio
110 / 150 min this week
Body
Sleep stages · HRV · BP
Nutrition
5 / 7 days tracked
Works with no signal. What you log offline syncs when you are back.
Your HRV dropped for nine days. Was it the training block, or what you started three weeks earlier?
Spectrum lays any five signals on one time axis and lets you shift one against the others by up to fourteen days, because a body doesn't answer on the same day you ask. Every pair is rebuilt from the days that actually carry a value on both series. Nothing is filled in.
Any signal against any signal · lag −14 to +14 days · correlation suppressed when there are too few real pairs to mean anything
Your weight is down four kilos. How much of that was fat?
Weight alone can't answer it, and neither can a scale that guesses. Body keeps every channel on one window (sleep and its efficiency, glucose, resting heart rate, composition by the method you actually use), so a change shows up next to everything that was happening around it.
Sleep, cardiac, metabolic, activity and composition channels on a shared window · four methods for body fat · fifteen circumference sites
You trained hard all month. Did the load land where you meant it to?
Sessions, volume and the shape of your week, with the zone-two debt read as debt, against the reference you set, not a target somebody chose for you. Strength, cross and cardio sit on equal rails, because your week has all three.
Strength, cross and cardio · every load number carries its published formula · thirteen-week frequency map · the full logbook
You hit your protein most days. What happened on the days you didn't?
Nutrition starts by telling you the truth about your own log: how many days you actually recorded, and the shape of the ones you skipped. Plan against real, intake against expenditure, hydration included. No projection, no 'if every day looked like today'.
Logged days audited against the window · plan versus actual per macro · intake versus expenditure · hydration with its own goal
Your training load fell for three weeks in April. Was it the programme, or was it what you were taking?
Alchemy keeps what you take with the half-life published for that molecule, so Vayl can model how much of it is still active on any given day instead of treating it as a date on a calendar.
That modelled curve lands on the same axis as your training load, your body composition and your blood panels. Which is the only way the question above stops being an opinion: you can see the weeks line up, or see that they don't.
Searchable by the name on the box, not just the chemical one · half-life from the published literature · supplements stay off the accumulation curve, because they don't belong on it
What the catalog holds
3,090
Medications
4,012
Substances in all
Three out of four are everyday medication
Half-life, as published
- Amlodipine40 h
- Sertraline26 h
- Atorvastatin14 h
- Losartan6.5 h
- Metformin6 h
- Prednisone3 h
- Ibuprofen2 h
- Amoxicillin1.3 h
Counted in the catalog on 2026-07-27
Vayl reads. The reading is yours.
It will not tell you what to take, what to eat or how to train. Not because it can't compute an opinion, but because the moment a tool hands you a verdict, you stop looking at the data that produced it.
Always
- Show the raw value, with the unit and the reference band
- Put what you took next to what your body did
- Break the line where data is missing
- Let you export every record you logged
Never
- Recommend anything at all
- Score you, rank you, or hand you a single number
- Guess a value you never measured
- Sell your data, or train anything on it
So, do you actually know what works for you?
Not what worked for someone on a forum, or what a study found in forty strangers. You. The answers are already sitting in the data you have been collecting for years, in four apps that never spoke to each other.
Vayl just puts it on one axis and gets out of the way.
Closed testing
The web terminal opens by invitation while we finish. Leave your address and you go in with the next batch.