Can AI build a medication reminder app like Medisafe?

Build a personal medication reminder web app to replace Medisafe Premium for my own use. I understand this is best suited to non-critical medications and that browser notifications are not fully reliable.

Tech rules:
- Plain HTML, CSS, and vanilla JavaScript in one file, plus a small service-worker.js so it can send notifications when installed to a phone home screen. No framework, build step, account, or server.
- Store all data in localStorage under one key, with a version number.

Data model:
- Medication: id, name, dosage, schedule (daily times [], weekly days+time, as-needed), pillsRemaining, refillThreshold.
- DoseEvent: id, medicationId, scheduledAt, status (taken, missed, snoozed, skipped), loggedAt.
- TaperSchedule: medicationId, changes [{date, newDosage}].

Features:
- Add a medication with a name, dosage, and schedule: fixed daily times, specific weekly days, or as-needed.
- At each scheduled time, if the app is open, show a prompt to mark the dose taken, snoozed, or skipped, each a distinct status.
- Refill countdown: track pills remaining and warn when it drops below a set threshold.
- Support a taper or dose-change schedule where the dosage itself changes on specific future dates.
- Prevent duplicate-dose logging: if a dose for the same scheduled time is already logged, a second tap asks for confirmation instead of silently creating a second entry.
- A printable adherence report for a date range: percentage of doses taken on time, missed, and skipped.
- On first load, show a clear one-time notice: this app is best for non-critical medications, and browser notifications on a locked phone are not guaranteed to fire, especially on iOS Safari.

Edge cases the build must handle correctly:
1. Missed-dose, snoozed-dose, and intentionally-skipped-dose are 3 distinct states, never merged into one.
2. Notifications requested when the phone is locked or the app tab is closed may not fire; the app must not claim or imply a dose reminder is guaranteed to reach you.
3. Multiple medications with different schedules (daily, weekly, as-needed) run independently without one schedule interfering with another's reminders.
4. Accidentally tapping "taken" twice for the same scheduled dose does not log 2 doses; the app asks for confirmation before logging a duplicate.
5. A refill countdown correctly decrements as doses are logged taken, and warns before the supply reaches zero, not after.
6. A taper or dose-change schedule (for example a steroid taper) correctly shows the new dosage on and after its effective date, not before.
7. A timezone change during travel shifts scheduled dose times consistently, and the app clearly shows the adjustment rather than silently changing dose times.
8. There is no caregiver alert or escalation feature; the app must not suggest it can notify anyone besides the person using it.

Backup:
- One-click "Export JSON" and "Export CSV" of medications and dose history.
- "Import JSON" restores after confirmation.

Layout:
- Mobile-first, large tap targets, readable at 375px wide, high contrast.

Exclusions:
- No publishing, App Store submission, accounts, payments, social features, drug-interaction checking, or caregiver/emergency escalation features. Do not present this app as a substitute for a clinical-grade reminder system.

Write a short README explaining where data is stored, how to export and restore it, how to install the page to a phone home screen for notifications, and a clear warning that this is not reliable enough for time-critical medications.

Keep it under 650 lines. This is a personal tool, not a product.

What is the short answer?

Yes, but with a real safety limit, not just a feature gap. AI can build a medication log with scheduling, refill tracking, and an adherence report to replace Medisafe Premium at roughly $4.99 a month in a few hours [VERIFY, since Medisafe’s own pricing page currently 404s].

The “but” matters more here than in most builds on this site. Medisafe’s free tier reportedly caps at tracking 2 medications since a change dated to January 2026 [VERIFY, no official Medisafe page confirms this]. That’s an annoyance for someone managing several prescriptions, and easy enough to solve with a personal build. The harder problem: browser and PWA push notifications are not fully reliable, especially on iOS Safari and when a phone is locked or the browser has been force-quit. For vitamins or supplements, an occasional missed notification is a minor inconvenience. For a time-critical prescription like insulin, a blood thinner, or an anti-seizure or psychiatric medication with strict timing, that notification gap is a genuine safety limitation, not something a good prompt fixes.

Option	Best for	Upfront cost	Ongoing cost	Main limitation
Build a rough prototype	Someone tracking 3 or more non-critical medications past the free-tier cap	$0 and 2–3 hours	$0	Notifications not guaranteed on a locked phone
Build a daily-use personal version	Someone who wants refill tracking, tapering, and an adherence report for non-critical meds	$0 and 6–8 hours	$0	Still no guaranteed notification delivery, no drug-interaction checking
Keep paying for Medisafe Premium	Someone managing time-critical medications who needs dependable alerts	$0	Around $4.99/mo or $39.99/yr [VERIFY]	Free tier reportedly caps at 2 medications [VERIFY]

The ready-to-paste starter prompt at the end of this page builds the daily-use version.

What should the first version include?

The first version needs the 3 dose states right and an honest warning about notifications, before anything else.

• Multiple medications with independent schedules
• Missed, snoozed, and skipped as 3 distinct states
• Refill countdown by pill count
• A one-time, visible notice that notifications aren’t guaranteed on a locked phone
• Deliberate exclusion: no drug-interaction checking, no caregiver alerts. Neither should be faked by a personal build; both need real clinical or multi-user infrastructure.

What can AI build reliably?

ChatGPT and Claude build the scheduling, logging, and refill countdown reliably. It’s a data model and a form, similar in shape to other tracker apps on this site.

Concrete things AI gets right:

• Independent schedules for multiple medications running without interference
• Missed, snoozed, and skipped logged as separate, distinct states
• A refill countdown that decrements correctly as doses are logged
• A printable adherence report by date range

Where will AI need human help?

This is the section to read carefully before you build. Reliable push notifications from a web app, especially on iOS Safari or when the phone is locked or the tab is closed, are a real limitation, not a bug to prompt away. A dose reminder can silently fail to reach you, and neither ChatGPT nor Claude can change how iOS Safari or Android background restrictions work. If you’re managing insulin, a blood thinner, or a psychiatric or anti-seizure medication with strict timing, do not rely on a self-built web app as your only reminder system. Use a purpose-built app with more dependable notification infrastructure, or pair this build with a physical backup like a pill organizer with its own alarm.

Duplicate-dose prevention needs your testing too. Confirm that accidentally tapping “taken” twice for the same scheduled dose doesn’t silently double-log it, since an accidental double-tap across devices is a specific edge case worth checking by hand.

The decision only you should make is which medications are appropriate for this build at all. Vitamins and supplements are a reasonable fit. Anything where a missed dose has real medical consequences is not, no matter how carefully the app is built.

There is no drug-interaction or dosage-safety database in a personal build, and there shouldn’t be one. That’s a real engineering and liability gap a DIY app can’t close, so don’t let the app’s confidence in its own UI convince you it’s checking anything it isn’t.

Can I get my data out of Medisafe?

Partially, with some uncertainty. Medisafe’s in-app Report feature under the Manage tab lets you pick a timeframe and medication and email yourself a CSV or Excel-style adherence report, and a separate PDF status report can be generated from the Measurements page [VERIFY, the official FAQ page could not be independently confirmed during this research].

The export trap: the export appears limited to adherence timestamps, taken, skipped, or snoozed, rather than a full account export. There’s no confirmation it includes dosage instructions, interaction warnings, or caregiver and family settings, and it’s unclear whether it’s gated behind Premium [VERIFY].

The practical path: export what you can, and manually re-enter your current medication list, dosages, and schedules into your new app rather than expecting a full history import. For a reminder app, the current medication list matters far more than historical adherence data.

How long will it take and what will it cost?

Rough prototype: 2–3 hours and $0. Scheduling, logging, and a basic refill countdown for a few medications.

Dependable daily-use version: 6–8 hours, including taper schedules, duplicate-dose prevention, and a printable adherence report. Build it on a free AI plan.

The math favors building if $4.99 a month matters to you and your medications are non-critical. It does not favor building if you’re managing a time-critical prescription: the cost of a missed notification there isn’t measured in dollars.

These are estimates from reviewing the feature set, not measurements from a completed build.

Which AI tool or approach should I use?

Claude or ChatGPT on a free plan handles the scheduling and logging build well.

Cursor isn’t necessary for a build this size.

Lovable doesn’t solve the core notification problem either, since the reliability gap is a platform limitation (iOS Safari, Android background restrictions), not something a hosted backend fixes.

What will I need to maintain?

Modest, with one important ongoing check.

Recurring work:

• Export a JSON or CSV backup every few months
• Periodically test that notifications still fire after phone OS updates, since permissions can reset
• Update medication lists and taper schedules as prescriptions change

Should I build it, buy it, or reduce scope?

Build it if your medications are non-critical, like vitamins or supplements, and you’ve accepted that notifications aren’t guaranteed.

Buy it if you’re managing a time-critical prescription. Medisafe’s notification infrastructure, while not perfect, is built and tested at a scale a personal project can’t match, and the stakes are too high to risk a silent failure.

Reduce scope if you just need a simple daily checklist without scheduling logic. A basic habit tracker or even a paper pill organizer with a physical alarm may serve you better and more reliably than any web app.

Recommendation: build it only for low-stakes medications where a missed reminder is an inconvenience, not a health risk. For insulin, blood thinners, anti-seizure or psychiatric medications with strict timing, keep a dedicated app or a physical backup system instead. If you’re also tracking daily habits or water intake, see the habit tracker build and the water tracker build.

What if I wanted to ship this to other people?

Shipping a medication reminder to other people raises real liability questions around missed doses and drug interactions, and it would need dependable, tested notification infrastructure across platforms plus caregiver and emergency escalation paths. That’s a clinical-grade product, not a personal project, and the safety bar is far higher than for any other app type on this site.

Starter build specification

• User: You, for your own use, tracking non-critical medications only.
• Problem: Wanting to track more than 2 medications and get reminders without paying Medisafe Premium, while accepting the notification reliability trade-off honestly.
• Core workflow: Add a medication and its schedule, get a reminder when the app can deliver one, log the dose status, check refill counts and the adherence report.
• Required features: Multiple independent schedules, missed/snoozed/skipped states, refill countdown, taper schedules, duplicate-dose prevention, printable adherence report.
• Deliberate exclusions: No drug-interaction checking, no caregiver alerts or escalation, no claim of guaranteed notification delivery.
• Data ownership: Stored in localStorage. Exported as JSON and CSV. Imported from JSON.
• Definition of done: The app clearly warns on first load that notifications aren’t guaranteed, 2 rapid taps on the same dose don’t double-log it, and a taper schedule shows the correct dosage on its effective date.

For scoping help, read can AI build an app for me.

Sources and verification

• Medisafe: Medication Management (App Store): pricing range across regions.
• Medisafe.com: product overview; the site’s own pricing page currently returns a 404.
• How Do I Export My Medication History in an Excel Spreadsheet? (Medisafe FAQ): export flow, not independently confirmed via fetch during this research.
• Medisafe Not Free: What to Do (pillo.care): third-party account of the January 2026 free-tier change.
• Medisafe No Longer Free: Caregiver Alternative (pillscircle.com): further third-party account of the free-tier cap.

This app was not built and tested for this page. The verdict is based on App Store listings and third-party comparison blogs, since Medisafe’s own pricing page was unavailable during this research pass. Several pricing and feature-gating details are marked [VERIFY]. Medisafe’s price was checked on September 24, 2026. This is not a substitute for a clinical-grade reminder system for time-critical medication.

Reviewed on September 24, 2026.