Building 3 SaaS Products as a Solo Dev with AI

My hobby is collecting hobbies.

I play several instruments. I've gotten into cars, archery, fitness, restoring vintage axes, and making beard oil. I keep chickens. I have six kids. My wife homeschools our teenager. I work full-time as a Director of Technology, and I do freelance AI consulting on the side.

I'm also building three SaaS products simultaneously.

Before you write me off as someone who doesn't sleep, I do. This isn't a hustle-culture story. The tools have just changed so much in the last couple of years that someone like me can actually pull this off without burning out.

Here's what I'm building, why, and what I've learned so far.

The three products

Every one of these started the same way: I had a problem, I couldn't find anything that solved it well, so I built something.

FlockNerd

FlockNerd is a chicken flock management app. Egg tracking, expenses, health logs.

Yes, it's niche. But if you raise chickens, you know that production numbers are one of the first warning signs that something is wrong. A sudden drop could mean a health issue, stress, or predators. I was tracking all of this in spreadsheets and it was terrible. So I built FlockNerd.

I originally built it from scratch to teach myself Next.js. It's been live for a while now, and recently I've been converting the progressive web app into a proper mobile app with Expo and React Native.

Here's the honest part: FlockNerd probably isn't going places as a business. I've put a lot of time and attention into it and I love using it, but I'm not under any illusions about the market size. Not everything has to be a unicorn. Some things you build because you love them, and that's enough.

Home Handoff

Home Handoff is a personal home guide for when someone else is watching your house. Housesitters, family, neighbors.

The idea hit me the last time we were out of town. I was digging through old Apple Notes trying to find the instructions I'd shared with someone last time. Half of them were wrong. I had videos from a year ago showing where I "used to" put dog food and how I "used to" water the chickens. Everything was scattered and outdated.

With six kids, multiple pets, and a house full of systems, "text me if you need anything" doesn't scale. I wanted one place to document all of it and share it with whoever needed it. Daily checklists, pet care instructions, which breaker to flip when the upstairs bathroom outlet dies. That kind of stuff.

There's basically nothing else like this for personal homes. Plenty of tools exist for Airbnb hosts and vacation rentals, but for regular people who just travel and need someone to watch the house? Zero direct competitors.

The validation was almost instant. Our housesitter used it once and told me I should sell it.

Homeschool Flow

Homeschool Flow is a homeschool management platform with separate parent and student portals.

My wife started homeschooling our teenager this past year. She kept telling me about different tools she'd tried, and they were all disappointing or frustrating. Every single one was parent-only. The student was completely left out. No dashboard, no way to see their own assignments, no way to submit work or track their own progress.

That's the differentiator. The student portal. Kids see their own dashboard, submit work, track what they've done and what's coming. There are 3.7 million homeschool students in the US, and every competitor I've looked at is dated or parent-only.

My wife encouraged me to keep working on it. She and my son use it every single day now, which is a kind of validation that's hard to beat.

Why three at once?

I get this question a lot, and the answer isn't some productivity hack. The economics of solo development just look completely different than they did two years ago.

A couple of years ago, what I could build was limited by what I knew, what my experience was, and what I had time to research and learn. If I wanted to add a feature that required a technology I hadn't used before, I was looking at hours or days of learning before I could even start.

Now I feel like I could literally build anything.

In fact, at work, we always joke that we can literally do anything except put page numbers on a PDF. (If you've fought with PDF generation libraries, you know.)

It also helps that all three products share a similar stack: Next.js, React, Tailwind, Clerk for auth, Vercel for hosting. Once you're fluent in that world, switching between products isn't that expensive. The patterns are the same.

The real bottleneck isn't code anymore. It's decisions and design taste. Knowing what to build and, more importantly, what not to build.

How AI actually fits into the workflow

I want to be specific here because there's a lot of vague "I use AI to code" content out there, and most of it is useless.

Here's my actual pattern:

I start with an idea. That can be something small, like a UI tweak, or something big, like ripping out an entire auth system. Then I plan it with an AI agent. Not "generate code." I talk through the idea. What are the consequences of this change? What's the scope? Why am I doing it? What haven't I thought through yet? The agent asks me questions. I ask it questions. We go back and forth until the plan is solid.

Out of that conversation comes a spec. What needs to change, acceptance criteria, sometimes specific instructions like "take a screenshot of the dev server after you make this change so I can see it." The spec is detailed enough that an agent can work through the implementation on its own. This is where Claude Code and OpenClaw come in.

Then I iterate. Something is always a little off. Maybe it doesn't look right on mobile. Maybe behavior in production differs from dev. I review, give feedback, the agent adjusts.

The biggest unlock is the infrastructure around all of this. I have CI/CD hooked up to my repos and OpenClaw connected to my dev box, which means I no longer have to be at my laptop to work on products. I can plan a feature from my phone while my kids are at practice. By the time I check back, the agent has a working implementation ready for review.

A typical build session is maybe two to three focused hours. In that window, with AI tooling, I can get done what used to take an entire weekend. That math is what makes three products possible. I'm not working three times as hard. The work just goes further than it used to.

Where AI helps the most: scaffolding, boilerplate, migrations, repetitive patterns. The stuff that used to eat an entire Saturday afternoon. When I rebuilt the FlockNerd mobile app in Expo, the agent handled most of the component migration while I focused on the screens that actually needed rethinking.

Where AI cannot help: product taste. Knowing your user. Deciding what to cut. Knowing that a feature sounds cool but will confuse people. Those are still human problems, and honestly, they're the fun ones.

What running three products has taught me

One good idea becomes three times the work. I'll have what feels like a really great idea. I'll work through it, implement it, make real progress in one app, and then realize I need to do something similar in another one. Sometimes one good idea turns into way more execution than I expected.

The tool-to-product gap caught me off guard. Now that I'm trying to get these out into the world, I'm realizing how much of a lift the marketing and polish pieces are. Going from a tool you use off and on to something you want to share with other people is a completely different kind of work. "Works for me" and "ready for strangers" are not the same thing.

Niche turns out to be an advantage. I'm not trying to compete with Notion or Asana. I'm building for chicken keepers, homeschool families, and homeowners who need a sitter. Small markets with real pain points and almost no competition.

Ship the ugly version. All three products are live right now, rough edges and all. You learn more from a real user clicking a janky button than from six more months in stealth mode. Home Handoff launched with a pretty minimal feature set. The housesitter didn't care. She just needed to know where the dog food was and what time to lock up. That was enough.

I also didn't do market research for any of these. No surveys, no focus groups. I built tools I needed for my own life, and it turned out other people needed them too. When you're your own first user, you know exactly what's annoying, exactly what's missing, and exactly when something is good enough.

Family isn't the tension — it's the source material

I think people expect me to talk about work-life balance like it's a negotiation. Like building products takes away from family time, and I have to carefully manage the tradeoff.

But my experiences are with my family. I'm fixing problems we have as a family.

FlockNerd exists because we keep chickens. Home Handoff exists because we travel and need someone to watch the house. Homeschool Flow exists because my wife is homeschooling our son.

I'm trying to build things that can supplement my income and give me more time with my family and provide for them. The goal isn't to work more. The goal is to build things that work for us, and if they work for other people too, even better.

And honestly? I get a lot more out of seeing other people use the tools that I create than I do from using them myself. Watching my wife and son use Homeschool Flow every day is more satisfying than any feature I could ship.

What's next

All three products are live. Real people are using them. Now the work shifts to polish, community building, and getting the word out, which, as I mentioned, is a completely different kind of work than building.

I'll be writing more about each product individually as I go. The specific decisions, the technical details, the stuff that went wrong. There's a post coming about the Homeschool Flow student portal that I'm particularly excited about, because the design decisions there were the hardest I've made across all three products.

If you're thinking about building something on the side, I'd tell you what I tell everyone: AI is making things possible right now that would've been way too time and energy intensive just a couple of years ago. If you have an idea and you think it's a good one, make it and see what happens. You can get feedback quickly and figure out what works and what doesn't. The cost of trying is lower than it's ever been.

And if your idea came from a problem your family has? Even better. You already have your first beta tester living under your roof.

---

Check out the products: FlockNerd · Home Handoff · Homeschool Flow

Follow along on X: @imjoshnewton

More at imjoshnewton.com