I just wanted to replace a bad app.
There was no clear product vision.
Also, there was no plan to build a residential automation hub.
The problem was much simpler: I had an IP camera at home and didn't like the experience of the app I had to use to access it.
It took a long time to open, it frequently froze, and at some points, the situation reached the point where I had to restart the camera itself in order to be able to follow the video again.
The camera was working.
The problem was everything that existed between her and me.
It was there that theNX Home.
The first problem was small
The first version had a very specific goal: to allow me to follow the camera directly from the browser.
Nothing more than that.
The streaming would continue to occur within my own network, but I would no longer depend on an application that was constantly in the way.
Appears to be a small change.
But she exactly solved the problem that bothered me.
Today I can follow the camera through the computer or cell phone, directly through the browser, with the stream happening almost in real time and without the instability I had previously.
The camera communicates with NX Home through go2rtc and the entire structure continues to run locally.
The most difficult thing, curiously, wasn't accessing the camera.
It was to make it work well on the web.
Browsers have restrictions that simply do not appear in the same way when we are working within a specific manufacturer's application.
Streaming to reach the interface without being blocked by the browser turned out to be one of the most laborious parts of the entire project.
But, after that started working, something more interesting happened.
The problem has changed.
The camera wasn't the only thing fragmented.
While the NX Home was maturing, I began to realize that the camera situation was part of a larger problem.
My home already had several smart devices.
Lights.
Air Conditioning.
TVs are a staple in many homes, providing endless hours of entertainment and connection to the world.
Lighting.
Echo Devices.
Each one of them worked.
But each one also tried to bring their own application, their own interface, and their own way of controlling something that, in the end, was part of the same house.
It was a connected home.
But the experience continued fragmented.
I had resolved this for the camera.
So I started questioning why not do the same with the rest.
It was at this moment that NX Home stopped being just a monitoring interface and started to become a hub.
Today, devices communicate through Home Assistant.
The NX Home has become the layer I really use to interact with this environment.
The difference appears to be architectural, but there is an important product change here.
I didn't want to build a new smart light bulb.
Also, I didn't want to replace the communication protocol of a TV or reinvent the residential automation ecosystem.
These parts already existed.
What I wanted to control was the experience.
Integração não é a mesma coisa que centralização.
There is an important difference between having multiple connected devices and having a connected system.
When each device demands a different experience, intelligence becomes scattered.
You need to remember where you control something.
Escolha o aplicativo certo para o seu negócio.
Qual fabricante é responsável por determinado equipamento.
Qual interface tem uma determinada configuração.
Technically, everything is connected.
In practice, you continue to act as the manual integrator of your own home.
Home Assistant resolveu uma parte importante desse problema para mim ao fornecer a camada de integração.
NX Home has shifted to solving another: how I want to see and use everything.
This distinction was important during construction.
Not every problem needs to be solved by creating a new infrastructure.
Sometimes, the best architectural decision is to simply use an existing layer and focus on what differentiates your product.
In my case, Home Assistant handles communication with most devices.
go2rtc resolves the streaming layer.
And NX Home begins to take responsibility for context.
After devices, came the person
At some point, another idea emerged.
Health.
At first glance, health seems like a completely different thing from residential automation.
And it probably would be if NX Home were still just a panel for controlling devices.
But it wasn't anymore.
I started seeing the house less as a set of equipment and more as the environment in which a significant part of my routine takes place.
It was then that Health began to make sense.
Today, the module already exists and allows for manual tracking.
The idea is to evolve it to integrate data from devices such as smartwatches and iPhone health services.
Workouts.
Routine.
Metrics.
Information that, isolated, would normally end up again scattered across other applications.
There is a curious line in this evolution.
First, I wanted to see a camera.
Then, control devices.
Then, understand the house better.
And now the system is also starting to look at the person who lives there.
Camera.
Devices.
House.
Person.
The scope increased, but not because I was simply adding features.
It increased because the definition of the problem was getting bigger.
Everything continues to be local
There is another decision that has become more important as the project grew: keeping the system running locally.
Today, the entire main structure of NX Home works within my own infrastructure.
This makes sense mainly when we're talking about something that starts to concentrate cameras, home devices, habits, and, in the future, health data.
The more context the system understands, the more important it becomes to think about where that context exists.
It's not just a performance decision.
It's also a decision about dependence and control.
A light shouldn't stop working because an external service is unavailable.
The camera doesn't need to send its stream to somewhere outside the network just so I can watch it inside my own home.
And extremely personal data doesn't necessarily need to start its journey on an external server.
Keeping everything local brings other responsibilities.
You need to take care of the infrastructure.
You need to think about availability.
Integrations can be more laborious.
The environment stops being completely abstracted by some provider.
There is a cost.
But in this project, it's a cost consistent with what I'm trying to build.
The next step isn't to put AI in everything
There is a layer that still hasn't been implemented in NX Home: AI.
And perhaps that's why I'm starting to think better about where it should enter.
The future idea isn't just to add a chatbot to the dashboard.
That would be the easy part.
The interesting thing would be to transform the context that the system already knows into something useful.
If NX Home understands the house, devices, certain usage patterns, and, in the future, part of my routine and health data, there is space for an assistant that really helps to interpret that environment.
Not just to answer questions.
But to perceive standards.
Understanding routine.
Suggesting adjustments.
Adapting behaviors.
Learning.
This may involve language models, machine learning, and other strategies.
But technology comes later.
Before it, there is a more important question:
what decision the system will be able to make better by understanding my context?
That is the part that interests me.
Because intelligence without context is just a new interface.
Building for you changes the way you discover the product
The NX Home is still evolving.
It will probably continue to change.
But there is something interesting in the path it has taken so far.
I did not start trying to predict the final product.
I started by solving a friction that existed within my own routine.
After solving one part, another became evident.
And another.
The camera showed the problem of applications.
The applications showed the problem of fragmentation.
Automation showed that the house could be treated as a system.
And the system began to reveal that the person was also part of that context.
Perhaps this is one of the most interesting parts of building tools for ourselves.
You don't need to imagine all the problems.
You live within them.
And, when you continue to observe after solving the first pain, you sometimes realize that what seemed to be the product was just the first layer of it.
The NX Home started as a better way to open a camera.
Today, I don't think much of it as an app for devices.
I think of it as a layer between my home, my routine, and the way I want to interact with the two.
The camera is still there.
It just stopped being the entire product.
It became the beginning of the story.
Share on:
No spam. Only content worth opening.