MARGIN

Inspiration

Most planning apps just track tasks and deadlines, they don't care if you actually have the time sleep or energy to get through what's on the list. I've had weeks where my schedule looked totally fine on paper and then just fell apart by Wednesday because it never accounted for how tired I actually was. I wanted to build something that plans around real capacity instead of some ideal version of the week that never actually happens.

What it does

MARGIN figures out how much time you realistically have in a week once sleep, commuting and existing responsibilities are taken into account, then helps you plan around that instead of an unrealistic schedule.

• Capacity Model — works out your real available time, not just whatever's left blank on a calendar
• Task Inbox — somewhere to dump tasks as they come to mind, you don't have to schedule them right away
• Week Lab — where tasks get placed against your available capacity so you can actually see what fits before committing to it
• Reset Engine — recalculates things when your week changes. missed a day, slept less than planned, need more hours than expected
• Calm Now — a quick way to step back when the week starts feeling like too much
• Support Card — short summary of where you're at for the week and what might need to shift around

How I built it

MARGIN is offline first, the capacity calculations and Reset Engine run locally in the browser so it doesn't need a server for the core stuff to work. No account needed, and planning data stays on your device where possible.

Challenges I ran into

Honestly figuring out how to model "capacity" in a way that felt real was the hardest part. It's pretty easy to make the math technically right and still have it feel completely disconnected from how people actually experience their time. I went back and forth on this more than I'd like to admit.

Accomplishments that I'm proud of

Getting the Reset Engine to recalculate cleanly was a big one, and keeping everything running locally without needing a server for the core experience. Small win but felt good once it clicked.

What I learned

Building this made it pretty obvious how much most planning tools just ignore real human limits. A lot of what makes planning actually work is being honest about your capacity, not just organizing your tasks better.

AI Tools Disclosure

I used AI mainly for debugging, code review, troubleshooting and improving my implementation during development.

The core Capacity model and project logic were coded by me. AI helped me find bugs, understand errors and suggest fixes.

For example, I wrote the capacity logic myself:

const availableHours = wakingHours - day.fixed - day.care - day.commute - recoveryReserve;

const adjustedDemand = day.flex * friction; const debt = Math.max(0, adjustedDemand - availableHours);

I tested and decided which fixes to use, while the underlying implementation and product decisions remained mine.

What's next for MARGIN

• Smarter capacity adjustments based on patterns over time
• Maybe cross device sync at some point, while still keeping data local first
• More personalization in how capacity actually gets calculated

Built With

• accessibility
• aes-gcm
• contextual-bandit
• css3
• csv
• github
• html5
• ics
• javascript
• json
• local-storage
• pbkdf2
• progressive-web-app
• service-worker
• web-crypto-api
• web-speech-api