I Built an AI App That Finds What You Can Build With Your Old LEGO Bricks

Muhammet — Sun, 26 Jul 2026 17:02:19 +0000

A few months ago, I was looking at a familiar scene that many parents probably know:

A huge box full of LEGO bricks.

Thousands of pieces.

Different colors, shapes, and sizes.

But the funny thing is… nobody knows what can actually be built with them anymore.

My kids would open the box, search for a few minutes, and eventually say:

"We don't have the right pieces."

Then they would ask for a new LEGO set.

That made me wonder:

What if the problem isn't that we don't have enough LEGO?

What if we just don't know what we already have?

That question became the starting point of Brickbeam.

The idea behind Brickbeam

The concept is simple:

Take a photo of your LEGO bricks
AI analyzes the pieces
Discover sets and creations you can build with your existing collection

Instead of buying another box of bricks, Brickbeam helps you rediscover the possibilities hidden inside the boxes you already own.

LEGO collections naturally grow over time. A child might start with one small set, then another, then another… and after a few years, hundreds or thousands of pieces end up mixed together.

Many of those pieces are forgotten.

Brickbeam is an attempt to solve that problem.

Why I built this

Like many indie projects, this started from a personal frustration.

I didn't start by looking for a market opportunity.

I started with a problem I personally experienced.

The interesting part about LEGO is that every collection is different.

Two families may own the same number of bricks, but their possibilities are completely different.

One person might have lots of wheels and plates.
Another might have many minifigure parts.
Another might have mostly architectural pieces.

A good experience needs to understand the user's actual collection.

That is where AI becomes interesting.

The technical challenge

At first, the idea sounded simple:

"Just recognize LEGO pieces from an image."

But very quickly I realized this was not a simple computer vision problem.

A real LEGO box is messy:

Pieces overlap each other
Similar parts have different variations
Colors change depending on lighting
Small pieces are difficult to identify
Some parts are only useful when combined with others

A child dumping a box of LEGO on the floor is basically a worst-case scenario for image recognition.

The challenge is not only:

"What pieces are here?"

The bigger question is:

"What can we do with these pieces?"

From recognition to creation

One of the most exciting directions for Brickbeam is moving beyond existing LEGO instructions.

Today, many LEGO apps focus on official sets and digital instructions.

But there is another interesting possibility:

What if AI could create new builds based on your personal collection?

Imagine:

"I want to build a spaceship."

AI looks at your available pieces and creates a design using only what you own.

Or:

"Give me a medieval castle using my bricks."

The goal would not be random AI images.

The goal would be real, buildable creations.

That introduces a completely different challenge:

Is the structure physically stable?
Are the required pieces available?
Are the building steps understandable?
Can a child actually build it?

This is where I think the future opportunity exists.

Building a product as an indie hacker

Building Brickbeam has also been a lesson in choosing what to build first.

The perfect version would probably include:

Complete LEGO inventory management
Advanced AI recognition
Custom AI-generated builds
Community sharing
Collection history
Recommendations based on previous creations

But as a solo developer, you need to start somewhere.

The first version focuses on the core question:

"Can we help people discover more value from the bricks they already have?"

Everything else can grow from there.

The moat question

A few people asked an important question:

"Could LEGO build this themselves?"

It's a fair question.

Large companies can build many things.

The challenge for an indie product is not always having a feature that nobody else can copy.

Sometimes the advantage comes from:

understanding a specific user problem deeply
moving faster
building a community
creating a personalized experience

For Brickbeam, the long-term vision is not just recognizing LEGO pieces.

It is becoming a creative companion for someone's LEGO collection.

An app that understands what you own and helps you create things you didn't know were possible.

What's next?

Brickbeam is still at the beginning.

There are many things I want to improve:

Better piece recognition
More accurate build suggestions
AI-generated custom creations
Better support for large collections

But the first step is already here.

Brickbeam is available on the App Store:

[App Store link]

If you have a LEGO collection (or kids who do), I would love to hear your thoughts:

Would you use an app like this?
What would make it more useful?
What would you want AI to help you build?

This is my first step building Brickbeam in public.

Let's see where it goes. 🚀

DEV Community: Muhammet