Schoolnight
A vibe-coded family app that turns Canvas homework into one calm daily summary
At a glance
I acted as design lead and executive sponsor for Schoolnight, a family product that reads a student’s Canvas account once a day and answers what’s done, what’s missing and what’s coming. I found the problem, researched the market, set the thesis, designed every surface and used Claude (AI) to help build it and its working prototype.
The challenge
Missing assignments were surfacing in progress reports weeks too late, and the nightly "Did you do your homework?" question was turning into an argument. Parents needed facts instead of interrogation, without logging into anything.
What I led
- Found the problem in my own family, then researched the market: the Canvas Parent app, parent forums and a 2025 homework-conflict study
- Defined the product thesis: the list becomes the bad guy, so the parent doesn’t have to
- Built a Canvas integration that surfaces hidden deadlines teachers keep in documents, handles time zones and works with parent observer accounts
- Designed a seven-step signup, a shared homework board, a daily email and a rewards engine with a parent approval gate
- Created a design system with K–5 and 6–12 age bands, per-school color theming and accessible light, dark and system themes
- Architected a fast, dependency-free build on AWS with CloudFormation, CloudFront and strict security headers
- Built a working interactive prototype of the email, board, rewards and setup
- Wrote honest release notes and mapped the trust work needed before any real student data
The outcome
In under a month, Schoolnight went from a personal homework board to a fully interactive v1.0 prototype with a marketing site, signup, multi-student board, daily email and pricing. The interaction design is settled, and the remaining work, the trust and privacy layer, is clearly mapped.
Overview
Schoolnight is for parents of Canvas students. It reads a student’s Canvas account once a day and answers three questions before a parent walks in the door: what’s done, what’s missing and what’s coming. It arrives as one email and one shared board, and nobody has to log into anything. It started as a tool for my own family in August 2026 and became a full v1.0 product in under a month.
How I led it: I acted as design lead and executive sponsor. I found the problem, researched the market, set the product thesis, designed every surface and the design system, and used Claude (AI) to help build the product and its working prototype.
What’s in v1.0:
Daily email: a fifteen-second read with a needs-attention band, points at risk and sections that hide when empty
Shared board: stat tiles, work grouped by urgency and project deadlines found in teachers’ documents
Rewards: stars for K–5 and streaks for 6–12, behind a parent approval gate
Setup: a seven-step signup, multi-student support and three plans
Trust: if Canvas can’t be read, the board stays untouched and a failure notice goes out
Services
Customer & Market Research, Product Strategy, UX/UI & Interaction Design, Design System, AI-Assisted Development, Interactive Prototyping, AWS Architecture
Client name
Schoolnight (my own product)
Team
John Karl, Design Lead & Executive Sponsor
Prototype
Market research
User research
UX Decision
Go-to-market
This is a working prototype. Follow the guided steps or switch tabs and use the controls. Everything responds live.
Market research
Goal:
Understand why families still fight about homework when the grades are already online.
Canvas Parent app: rated 4.4 from about 111,000 App Store ratings. It shows grades, but parents must find and set up missing-assignment alerts themselves, a question that keeps coming up in the Canvas Community
The gap: existing tools report data. None of them change the conversation at home
User research
Goal:
Ground every decision in real input from the people who use it, and say plainly what still needs validating.
My own family (real input): missing assignments surfaced in progress reports weeks too late, and the nightly “Did you do your homework?” turned into an argument
Academic research: a 2025 study of 78 families and 602 recorded homework sessions found seven recurring conflict types, and even behaviors parents meant as helpful were linked to conflict
Proto-personas (hypotheses): the working parent who checks once a day, the co-parenting household and the student who wants to be trusted
Research plan: a two-week diary study with families, a fifteen-second read test of the email and a district pilot before any real student data
Why Each UX Decision Was Made
Goal:
A parent should get the whole picture in fifteen seconds, without starting an argument.
Facts instead of interrogation: one shared board both people look at, because the research shows well-meant questions still cause conflict
No student checkbox: a box you tick yourself is a claim, not evidence, so only what Canvas recorded counts
A parent approval gate for rewards: points are awarded only after a parent reviews them
Lead with points at risk: missing work shows what can still be recovered, not just what went wrong
Polite by default: contact-the-teacher drafts stay non-accusatory
Stale data never wears today’s date: a failed Canvas read leaves the board untouched and sends a notice
Age bands and school colors: K–5 and 6–12 versions with per-school theming and accessible light and dark modes
Go-to-Market: Pricing, Positioning and Trust
Goal:
Launch with a clear price, an honest promise and the trust work mapped before real student data.
Positioning: “The list becomes the bad guy, so the parent doesn’t have to”
Pricing: Free, Family at $7 a month and School, positioned against the cost of one tutoring hour
Marketing site and FAQ: the FAQ answers the uncomfortable questions first, like “Is this spying on my kid?”
Channel: Canvas works the same across districts, so a daily digest travels to any Canvas school
Trust before launch: server-side sessions, per-district Canvas agreements, separate parent and student views, a real backend and a FERPA and COPPA legal review
Schoolnight marketing site and sample daily summary.
Final observations
Schoolnight shows how I lead a product from a real problem: start with the people involved, ground the thesis in research, design for the relationship and not just the data, and be honest about what isn’t done yet. The interaction design is settled. The next steps are the diary study, a district pilot and the privacy and trust work that has to be in place before any real student data touches it.
John Karl
19 Years • Product & UX Design Leader Leading Mobile Product for 500K+ Users
If you like what you see or have any questions, feel free to send me an email anytime.
Selected work
[2010 -2025]
