by Kulsum Shannan
Full disclosure: I currently intern at Perfai Security, the authorization and access control security testing platform mentioned later in this article. This isn't a sponsored post — I genuinely built this app as a personal project, and because I work at Perfai, I naturally used it as part of my release process. If you think either the app or Perfai could be improved, I'd genuinely appreciate your honest feedback.
Three Problems It Solves
Forgotten Home Maintenance
Most homeowners don't intentionally neglect maintenance — they simply forget. HomeKeeper automatically reminds you about recurring maintenance like replacing HVAC filters, cleaning dryer vents, testing smoke detectors, servicing appliances, and seasonal home care, before those tasks become expensive problems.
Household Knowledge Lives in One Person's Head
Every household has one person who somehow remembers everything: when the dishwasher was serviced, where the appliance manuals are, which contractor fixed the plumbing, what paint color was used. When that information isn't documented, everyone else starts from scratch. HomeKeeper turns that knowledge into something the entire household can safely access.
Home Information Is Scattered Everywhere
Receipts live in email. Warranty PDFs are buried in Downloads. Maintenance photos are somewhere in the camera roll. Calendar reminders disappear. HomeKeeper keeps everything connected — recurring schedules, maintenance history, appliance information, documents, photos, notes, and reminders, all in one place.
The App
HomeKeeper OS is a smart home maintenance and cleaning management platform that helps homeowners and families keep track of recurring chores, appliance care, repairs, and important home documents. The app provides automated reminders for tasks like replacing filters, cleaning appliances, seasonal maintenance, and inspections while creating a long-term history of everything done in a home.
Key Features:
- Recurring cleaning and maintenance schedules
- Smart notifications and reminders
- Appliance tracking with warranties, manuals, and service history
- Room-based organization
- Maintenance history timeline
- Task assignment and collaboration
- Document and photo storage
- Dashboard analytics for home upkeep
Roles:
- Household Owner: Full control over the household, members, settings, and all data
- Household Admin: Manages tasks, schedules, rooms, and maintenance records
- Member: Completes assigned tasks, adds updates, and uploads photos
- Guest: View-only access to shared information
The app is designed to become a digital memory for your home, helping users remember important maintenance tasks and preserve household knowledge over time.
Why I Built HomeKeeper OS
I started thinking about all the maintenance tasks people constantly forget around the house — not daily chores like taking out the trash, but the things that happen every few months, or even once a year. When did I last replace my HVAC filter? Have I cleaned the dryer vent recently? When should I flush the water heater? Did I ever replace the smoke detector batteries? Which paint color did we use in the guest bedroom? Where's the warranty for the dishwasher?
None of these are difficult. The problem is remembering them. Every reminder ends up in a different place — Notes, Google Calendar, sticky notes, emails, or simply forgotten.
I wanted one application that would become the memory of a home. Not another to-do list, but a system that remembers maintenance history, schedules recurring tasks automatically, stores manuals and warranty documents, keeps track of appliances, and lets family members collaborate without giving everyone full control. That's what became HomeKeeper OS.
How I Built It
Like a lot of people lately, I wanted to see how far modern AI development tools could go. Instead of building the application page by page, I started with a single detailed prompt inside Lovable. The prompt described the overall product vision, user workflows, a multi-tenant architecture, role-based permissions, recurring maintenance schedules, notification logic, dashboards, authentication, PostgreSQL database structure, Row Level Security, API endpoints, file uploads, and responsive UI.
From that single prompt, Lovable generated a surprisingly complete full-stack application. I then iterated on the generated code to refine workflows, improve the user experience, connect integrations, polish the interface, and make the authorization model behave consistently throughout the application.
By the end, HomeKeeper supported authentication, multiple households, Household Owners, Admins, Members, and Guests, recurring maintenance schedules, smart notifications, appliance management, document uploads, maintenance history, role-specific dashboards, and a multi-tenant PostgreSQL architecture. It looked and behaved like a real SaaS product.
The Part AI Builders Don't Solve
This post isn't really about HomeKeeper. It's about something I think many developers overlook.
Today it's incredibly easy to build software that looks production-ready. Modern AI builders like Lovable, Bolt, and v0 can generate polished dashboards, authentication flows, CRUD operations, and beautiful interfaces in hours instead of weeks. But none of that proves the application is actually secure.
A polished UI tells you nothing about whether someone can access another household's maintenance history, download documents they shouldn't see, modify someone else's recurring tasks, bypass role restrictions by calling the API directly, or access another tenant's data. Those problems don't usually appear during normal testing. Everything looks fine — until someone intentionally tries to break it.
Why I Tested It With Perfai
Since I intern at Perfai Security, using it as part of my release process was the obvious choice. I'm mentioning it because it's genuinely what I used, not because I'm trying to pretend I'm an unbiased third party.
After deploying HomeKeeper, I ran it through Perfai to test the live application. The goal wasn't to review my code — it was to test whether the authorization model actually held up against direct requests to the deployed app. That meant checking things like tenant isolation, role-based access control, ownership validation, API authorization, file permissions, broken access control, privilege escalation, and IDOR vulnerabilities. Those are exactly the kinds of problems that are difficult to notice by clicking around the UI.
The first scan came back with 5 critical access control issues — nothing that showed up in normal use, but real gaps that a determined user could have exploited: a couple of endpoints that didn't fully check household membership before returning data, and a role check that wasn't enforced consistently on the backend. I fixed each one directly in Lovable and reran the same scan. The second report came back clean, with zero issues remaining.
Whether you use Perfai or another tool, I think this kind of testing should become a normal part of shipping AI-generated applications. Building software quickly is becoming easy. Building software that's safe to hand to real users is still a separate job.
I'd Really Like Your Feedback
Developers are usually the toughest audience, and that's exactly why I'm sharing this here. I'd genuinely appreciate feedback in the comment section on two things:
HomeKeeper OS [https://homebase-recall.lovable.app] — Is this an application you'd actually use? What features feel unnecessary? What would make it genuinely useful over the long term?
Perfai [https://perfai.ai/] — Since I work there, I'm especially interested in hearing honest opinions. If the testing workflow, reports, or developer experience could be better, I'd love to hear it. Skeptical feedback is far more valuable than polite agreement.
I don't expect everyone to agree with my conclusions, but I'm hoping this project starts a conversation about what the "last mile" of AI-generated software should look like. Building an app is no longer the hard part. Knowing it's ready for real users still is.
Top comments (0)