He Inventoried His Fridge. Then Built an App That Did It for Him.

There’s something electric about seeing an idea go from thought to thing.

That’s what It Exists is all about, real stories of people who stopped overthinking and started building.

Some found their spark in NTE Pro (our library of 4,000+ startup ideas). Others want to use NTE Zero to One to get hands-on help bringing their MVP to life.

No gatekeeping. No guesswork. Just momentum.

Here’s your dose of proof that building is the best way forward.

Let’s dive in.

By day, Nick Markman is VP of Product at a real startup.
By night, he’s using AI to get his groceries under control.

The result?
MealMuse, an AI-powered web app that turns photos, receipts, or even your voice into personalized recipes, complete with a shopping list and pantry tracking.

Think of it like having a private chef, nutritionist, and fridge-whisperer in your pocket.

And he built the first working version… in 24 hours.

It all started with a fridge, a spreadsheet, and a $400 grocery bill.

The Grocery Bill That Broke the Camel’s Back

Nick and his wife love to cook. But month after month, the grocery bill made zero sense.

“We’re two people. Why are we spending this much?”

The problem wasn’t the food. It was the system:
They’d grab a recipe, run to the store, and buy every ingredient from scratch, even if half of it was already at home.

Limp produce. Duplicate cans. Pantry items going stale.
It was wasteful, time-consuming, and expensive.

So Nick did what any decent product manager might:
He opened up ChatGPT, listed out everything in the fridge, and asked for recipes.

At first, the results were “meh.”
But as the models improved, the meals got better.
They cooked more. Wasted less. Grocery bills dropped 30%.

And one day, he looked at his little workflow and thought:

“Wait… this could be an app.”

From Hacky Workflow to Working App in 24 Hours

In March, Nick signed up for a Lovable AI hackathon.

No grand plan. No pitch deck.
Just a personal system that worked, and a Saturday to build it.

He dropped in some ChatGPT-written PRDs, used Lovable to spin up the frontend, Supabase for the backend, and got to work.

By the end of the weekend, MealMuse was live.

Upload a photo, receipt, or voice note.

MealMuse auto-detects ingredients, applies your dietary preferences, and generates recipes using what you already have plus a smart shopping list for the rest.

It was rough. It was hacked together.
And it worked.

Total build time: 24 hours
Total cost: ~$15

“It wasn’t pretty. But it did the job.”

Lesson 1: Build the Tool That Works for One Person - Then Share It

Nick wasn’t chasing a billion-dollar TAM.
He was chasing a 30% savings on groceries.

There was no launch plan. Just a janky workflow that happened to work really well for him and his wife.

“I could’ve kept doing it manually. But I had a hunch I wasn’t the only one with this problem.”

That’s the unlock.

You don’t need a startup idea.
You need a working system that makes life better for one person, even if that person is just you.

Start niche. Start personal. Start real.
Then let it grow.

The Feature He Almost Skipped

One of the hackathon rules was: use a speech-to-text API.
Nick hadn’t planned on it. But to check the box, he added a voice input.

He spoke:

“Half a red onion, two eggs, a block of cheddar…”

MealMuse transcribed it, extracted the ingredients, and knew cheddar was dairy and didn’t judge you for it.

It worked so well it became his favorite part and users felt the same. He even added cheeky validation responses if someone tried uploading a selfie or mumbling nonsense.

The result wasn’t just functional. It was fun.

Lesson 2: Constraints Can Create the Best Surprises

That feature? It only exists because of a random rule.

“If that requirement wasn’t there, I wouldn’t have built it. But it turned out to be the most delightful part.”

Sometimes the best UX lives in the ideas you almost didn’t build.

Follow the constraint. Then find the magic inside it.

Users Showed Up. Nick Showed Up Harder.

A few weeks later, Lenny’s Newsletter featured MealMuse.

The response was instant:

• 1,400+ visits

• 100+ free accounts

• 20+ people built full meal plans

• 100+ recipes generated

And then the messages started:

“I want you to succeed… because I need this to exist.”

So Nick did what good builders do:
He followed the pull.

He messaged users. Fixed bugs. Improved the recipe flow.
Cleaned up the prompts. Smoothed the edges.

“Not for growth. Not for metrics. Just because people gave a damn.”

Lesson 3: Follow the Pull

Nick’s not quitting his job.
He’s not raising money.
He’s not making “MealMuse 2.0 with blockchain integrations.”

But when people start showing up, unprompted, unmarketed and tell you your thing matters?

That’s not noise. That’s signal.

“I’m still figuring out how far to take this. But I’ve seen enough to not stop.”

You don’t need a roadmap.

You just need a reason to keep going.

When users care, that’s your cue.

Follow the pull.

Bonus Round: He’s a VP of Product. But This Felt Different.

Nick builds product all day.
He knows how to write specs, run sprints, manage engineers.

But MealMuse?

This was different.

No Jira. No meetings. No grooming sessions.
Just a stack of sticky notes and a long Saturday.

“I scrapped the Notion board and just wrote bugs on stickies. Finished the whole pad by the end.”

It was messy. Fun. Immediate.

The pure joy of building something useful, fast.

What’s Next for Nick

Nick is still tinkering with MealMuse, improving the recipe flow, talking to users, and seeing where the energy leads.

He’s also exploring new builds in the agentic AI space.
More tools. More speed. More experiments.

And when we asked what he thinks needs to exist?

“I’d love to see AI-powered tools that radically improve education. Something that makes personalized learning fun and accessible, especially for K–12. One-on-one tutoring shouldn’t be a luxury.”

That’s a callout to the builders.

What You Can Learn From Nick

🛠 Start with your own life.
If it fixes your day, it’s worth building.

🎯 Delight hides in constraints.
That voice feature? A checkbox requirement. Now it’s the magic.

📬 Follow the pull.
Real signal sounds like: “I want this to exist.”

🧠 You don’t need to be an engineer. You need clarity.
Nick used Lovable, Supabase, Cursor, Claude. The magic wasn’t the stack, it was the plan.

🚀 AI is your team now.
Use it like a PM. Guide it. Lead it. Let it work while you think.

Try MealMuse or Follow Nick
👉 Try MealMuse — the AI-powered kitchen assistant
👉 nickmarkman.com — see more projects
👉 Nick on LinkedIn — reach out

Nick didn’t wait for permission.
He had a fridge full of groceries and a spreadsheet that got out of hand.
24 hours later, he shipped MealMuse, a personal chef, nutritionist, and meal planner in one.

Here’s the truth:

You don’t need funding.
You don’t need 10,000 followers.
You need to build something one person actually wants.

That’s how it starts.

Not sure what to build?

🔥 NTE Pro gives you 4,000+ real startup ideas, complete with GTM plans and monetization paths
⚒️ NTE Zero to One pairs you with us to help turn one into a working product. Fast.

You’ve got time this weekend.
You’ve got AI tools.
Let’s go.

Let’s make it exist.