What else could be doing this?
You started something, and something changed. Repeats reads your medicines, supplements and food together, and tells you what each has actually been linked to, and how strong that link really is.
It never tells you to stop anything. It gives you a question to take to your prescriber, with the citation attached.
Try the lookup
free · no account
Running against a fixed sample shelf, alphabetically, exactly as the app does. It will not tell you which one it is, because it does not know and neither does anything else that claims to.
Australian medicines, Australian crisis lines, Australian labelling rules
Everything it does, and the reason for each one
Scroll, or pick a step. The phone is the real thing, not a screenshot.
Everything on one shelf
Prescriptions, supplements, food, the lot. Read together, because that is where the interactions live and nobody else is looking at them together.
Complementary medicines sit beside prescriptions
St John's Wort next to an SSRI is the case this was designed around. A pharmacy shelf and a health-food shelf are the same shelf as far as your body is concerned.
Every claim carries its tier
Six tiers plus a refusal, and a direction-of-causation reading on every one. A card without a direction does not render, because most of these links run both ways.
Numbers, never the word
13 in 100 and 87 in 100, both at the same size. Pairing "rare" with its number pushes the estimate up rather than down. In the sharpest test, from a stated 0.04% to an estimated 18.0%.
You bring the symptom
Type what you are getting; it matches against labelled effects and stops. Alphabetical, never ranked, and it never names a condition.
One output shape, always
A question for the person who prescribed it, with the sources attached and dated to the label version. Never an action about a prescribed medicine: not stopping, not a dose, not which of three is most likely.
Five things a day, none of them scored
Deliberately not PHQ-9 or DASS-21, because administering a scored instrument is screening. No streak, no completion percentage, and skipping costs you nothing.
Four decisions, and what they cost
The word "rare" is gone
Product information sorts side effects into bands. We used to show the band and the number, because that felt obviously more informative. It is the worst of the three options that have been tested: adding the word pushes the estimate of "any side effect" from 18.7% to 31.1%.
Both halves now render at the same size, and positive framing cuts symptom attribution from 54.5% to 39.2% without reducing how many symptoms people have.
Insomnia
sertralineWithdrawn is the tier that matters
An alarm raised and later retracted by the body that raised it. Most sources never update, so the retraction never travels as far as the alarm did.
We do not get to assume a correction lands, either: across 205 effects and 60,861 participants the mean correction effect for science claims was not significant. So nothing is marked read, and no screen counts corrections as though each were a mind changed. The old worry that correcting a belief entrenches it is separately dead: 52 claims, more than 10,000 people, no backfire.
Montelukast
10 mgThe check-in lost half its questions
Eight became five. Water intake went because the within-person evidence is not there. Hours slept went for a different reason: it does have evidence, at roughly 0.04–0.05 SD on next-day affect, but sleep quality beats it and we were asking both.
Cut for redundancy is a different finding from cut for absence, and we had been reporting the first as the second. Nothing here is summed or compared to a cut-off, and a 2025 review found no robust evidence that mood tracking alone treats anything.
Today
3 FebHelp is siloed from everything else
One tap from every screen, unconditional, and never triggered by what you typed. Routing you on your symptoms is triage, and triage is what makes the whole app a Class III medical device.
Nothing from your shelf renders on it; a shelf is a means inventory. That silo is a precautionary call with no empirical data behind it, and we would rather say so. What does have data is safety planning delivered by a clinician: RR 0.57, NNT 16. No unguided digital version has been shown to carry that, which is why the app does not build you one.
Help, right now
Three people, one product
You are managing something
You take more than one thing, something changed, and you want to know what is worth raising without being told what to do about it.
You just want your scripts somewhere
Share the eScript token you already got by SMS and it lands on a shelf. If that is all you ever use it for, that is a completely reasonable way to use it.
You do not believe us
Good. Every number on this site opens its source, its quote and our own tier for it. The research is published unedited, including the parts that contradict us. Start there →
Put one thing on the shelf and see what it says.
Free to download, free to read. No account needed to read anything.