Fun fact: this will be the first year that both of my kiddos (ages 5 and 12) will be home all summer while I continue working full-time, also from home.
Fortunately, one of my super powers is overthinking problems to the point it becomes a hyperfixation!
My Solution: Build a Self-Hosted Web App with Claude Code
With all this talk of AI and vibe coding, I had already been considering creating an app of some sort to experience the workflow for myself. Once I started brainstorming what our summer could look like, I came to the realization that a summer schedule app might be the right solution and fulfill both needs.
Here were my original goals for the app:
- Find a way to balance fun activities, necessary boredom and screen time.
- Have the app sync events added to our family calendar.
- Have a one-click "swim day" option that re-arranges their daily schedules so we can go swimming in the afternoon.
- Color code everything, of course.
- Save all changes locally to a database so updates to the app wouldn't overwrite pre-planned schedules.
- Keep the app self-hosted so it's only available to us at home over WiFi.
I then proceeded to set up Claude Code, a local Python environment, a new Google Cloud Project under my personal account, and an API key for Google Calendar so that I could sync schedules.
Then I wanted a bit more, and the complexity evolved to include:
- Ability to drag-and-drop activities that didn't sync from the calendar.
- A bonus "chore" section where the kiddos could earn extra screen time once complete. (I gave into their puppy dog eyes after explaining the screen time limits, ok?)
- A new "planning" page where I could add and balance special events between both kiddos.
- A new "bucket list" page where we could check off overall goals for the summer.
I found the output wasn't perfect. At one point, the app stopped synching Google Calendar correctly and would mix-and-match dates (even though it had been working in a previous iteration.) I also used 10% of our PRO plan on this one project. ðŽ But hey, I now have a full summer schedule app that our little family can use to stay more organized - hopefully.
What I Learned During My First Vibe Coding Project
#1: Logic is Everything
Back in college during our introductory programming classes, our professors had us write out logic systems with pencil on paper. We'd start the project writing our requirements in plain English, then move to pseudo code. (Pseudo code is exactly like it sounds: part plain English and part code - just enough to help you keep track of variables, conditions and whatnot.)
I will always appreciate that they forced us to do this because it has made me a better developer. I understand how a system needs to be set up and can explain how tasks will need to be completed without code - which is really what vibe coding is.
#2: It Helps to Know the Terminology
Being in the industry and having a formal education around programming in general helped me speed up "conversations" with Claude in order to iterate on the app. I knew what to do in order to set up a local instance of GitHub, generate an API key, install Python via command line, as well as some other minor tasks. BUT I found it wasn't necessary. A novice could easily have a separate Claude chat open to clarify requests in order to move to the next step.
#3: It's Way Too Easy to Skip Security
I didn't proof the code ð I know it's a big no-no, but I wanted to see what would happen if I skipped concerns over security altogether. I clicked through every security-based prompt about exposing API credentials just to see what would happen.
I have since cleaned things up, of course, but don't be me and skip security entirely. If you're unfamiliar, here's a super quick list of things to check for:
- Encrypt user data, especially usernames/emails and passwords. Don't store it as plain text.
- Secure your API credentials so they're not public facing, otherwise people might hijack them costing you a lot of money.
- Sanitize and validate all user inputs to prevent SQL injection and XSS (cross-site scripting) attacks. If you're not entirely sure how to do this, prompt your AI tool to for you.
- Back up your data somewhere separate from the server itself.
- Make sure your self-hosted server isn't exposing unnecessary ports; only what the app needs should be open.
#4: It Gave Me Fresh Insight Into Scope Creep
Each time I thought I felt content with the app and started playing around with it, I would then want to add another feature, and another, and another. I'll probably keep adding more, honestly.
But, this gave me renewed insight into how quickly the scope can change to become more complex. First, it was changing some colors. Then, it was changing the interface to accommodate dynamic color options. It's fair to say I likely doubled - if not tripled - the original scope of my app.
I always have a hard time telling clients something is out of scope because I want them to be happy with the service they receive. But, I also know that, as a developer, making minor change after minor change can consume a lot of extra time that wasn't originally accounted for in the project. Having this fresh insight is helping to refine some of our processes and how we write proposals.
#5: Vibe Coding is Fun as Hell
I yapped about the app for a solid 30 minutes to my husband, then showed it off to my oldest to get his feedback. (Lesson learned, if you mention screen time limits, they'll deviate from the original conversation to figure out how to increase those limits.)
What This Means for Small Businesses Who Want to Try Vibe Coding
I've been building websites since 2006. As someone who codes for a living, I expected to breeze through this project to a degree. However, I skipped security prompts like a rookie, tripled my own scope, and spent too much time picking apart tiny details that didn't contribute to my goals. (Whoops!)
And that's kind of the point. Vibe coding lowers the barrier to development, but lower barriers don't eliminate the pitfalls. They just make them easier to stumble into without realizing it.
We know our small business and nonprofit friends will be tempted to vibe code a new tool or internal app at some point (if you haven't already). Just remember to think through your logic before you start prompting, don't skip the security steps, and always define your scope before you get started.
Ultimately, my hyperfixation superpower came through. I over-engineered a summer schedule, learned some things along the way, and now have a fully functional app that my kids will probably ignore before the end of June. Worth it!