Nothing Just Let You Build Your Own Apps by Talking to Your Phone. Is This the Future or a Party Trick?

Full disclosure: celltech repairs Nothing phones. This article was written independently.
The demo starts simply enough. A Nothing employee types a sentence into a text field: "I want a tea timer that screams when the tea is ready." Thirty seconds later, a widget appears on the homescreen. It has a start button, a countdown, and — yes — it screams.
This is Essential Apps, the feature Nothing Technology launched in beta today for Phone (3) owners. The pitch is deceptively straightforward: describe the app you want, and your phone builds it. The tech industry has already given this a name. They are calling it "vibe coding."
If that sounds like every AI demo you have seen in the past two years, I understand the fatigue. But after reading through every major reviewer's hands-on assessment this morning, and spending time understanding what Nothing is actually building here, I think there is enough substance to take seriously — if not quite enough to celebrate.
What Essential Apps Actually Are
Let us be specific about what this means, because Nothing's marketing language and the reality are not quite the same thing.
Essential Apps is a feature built into Nothing's "Playground" platform. You open it, type a natural language description of something you want on your homescreen — a habit tracker, a countdown timer, a currency converter, a weather widget — and an AI model generates functional code that creates it. The result appears as a widget on your homescreen. You can customise it further with additional prompts, share it with the Nothing community, download widgets other people have made, and remix them.
Think of it like this: instead of searching someone else's app store for something that roughly matches what you need, you describe exactly what you want and the phone builds it to specification. The industry calls this "vibe coding" — you set the vibe, the AI writes the code.
What Essential Apps are not — and this distinction matters — is a full app platform. These are homescreen widgets. They do not launch into full-screen experiences. They cannot access your phone's deeper functions. The Verge put it bluntly: "Essential Widgets would be a more honest name." That is a fair criticism, and Nothing would be wise to listen to it.
Why This Matters More Than It Sounds
Here is the industry context that makes a screaming tea timer worth writing about.
For fifteen years, the way we have used smartphones has followed exactly the same pattern. You want to do something, you go to an app store, you search, you download someone else's version of the thing you wanted, you live with its compromises. Apple and Google have trained billions of people to be consumers of software, not creators of it.
Nothing is asking a genuinely different question: what if the phone adapted to you, instead of the other way around?
This is not a new idea in tech circles. The concept of "generative user interfaces" — software that builds itself around what you need in the moment — has been circulating in AI research labs for years. But Nothing is, as far as anyone can tell, the first smartphone manufacturer to ship it as a consumer feature. Not as a concept video. Not as a developer preview. As a thing you can try on your phone today.
That alone is worth paying attention to.
The Company Behind the Ambition
To understand Essential Apps, you need to understand Nothing. And to understand Nothing, you need to understand Carl Pei.
Pei was born in Beijing in 1989, grew up in Sweden, and co-founded OnePlus in 2013 — the phone brand that proved you could sell flagship specs at mid-range prices and build a cult following doing it. He left OnePlus on his 31st birthday in October 2020, and founded Nothing immediately.
The pitch was deliberate: make tech fun again. Not faster. Not more powerful. Fun. The company launched from London, hired the Stockholm audio firm Teenage Engineering as design partner, and put out earbuds with transparent cases before anyone had seen a phone. When the Phone (1) arrived in July 2022, with its distinctive transparent back and LED "Glyph Interface," it was clear Pei was not building another generic Android handset. He was building a brand.
The numbers suggest it is working. Nothing has sold more than seven million devices globally. Lifetime revenue has crossed one billion dollars. In India, the brand grew 577 per cent year-on-year in 2024 — the fastest-growing smartphone brand in the country for five consecutive quarters. In September 2025, Nothing raised $200 million at a $1.3 billion valuation, specifically earmarked for what the company calls its "AI-native platform."
That platform is what launched today.
The Essential Acquisition
There is a piece of history embedded in the name that most coverage has glossed over. "Essential" is not just a marketing word. Nothing acquired Essential — the company founded by Andy Rubin, the man who created Android itself. Rubin's Essential Phone was a commercial failure but a design landmark: the first phone with a titanium frame, a near-bezel-less display, and a magnetic accessory system. The company folded, Nothing bought the brand and its intellectual property, and now the name lives on as the umbrella for all of Nothing's AI features.
There is something almost poetic about it. The creator of Android's company, reborn as the platform trying to move beyond the app model Android popularised.
How It Works: Playground and the Prompt
The technical architecture of Essential Apps is straightforward, at least in concept. You open the Playground app on a Nothing Phone (3). You type a prompt — plain English, no code, no technical knowledge required. The AI parses your request and generates widget code. The widget appears on your homescreen. You can iterate on it with follow-up prompts: "make the background darker," "add a reset button," "show the total in pounds instead of dollars."
The community layer is where it gets more interesting. Widgets you create can be published to Nothing's community platform, where other users can download them, rate them, and — crucially — remix them. Found a budget tracker that is close to what you want but not quite right? Fork it, modify the prompt, make it yours.
Under the hood, Nothing is using a large language model to generate functional widget code from natural language descriptions. The company has not disclosed which model or models power the system, but the $200 million AI investment suggests this is not a thin wrapper around someone else's API. Nothing appears to be building — or at minimum, heavily fine-tuning — something purpose-built.
What the Reviewers Actually Found
Every major Android publication got hands-on time with Essential Apps today. Here is what they reported, presented fairly.
| Publication | Verdict | Key Finding |
|---|---|---|
| The Verge | Conflicted | "Fun, until you try to make them useful" — buys the vision, questions the execution |
| Android Authority | Surprisingly capable | Built the screaming tea timer; found rough edges but genuine potential |
| 9to5Google | Cautiously interested | "Vibe code your homescreen" — intrigued but "not fully sold" |
| Android Central | Enthusiastic | "Exactly what I need" — saw practical, everyday potential |
| Heise (Germany) | Strategic view | Called it the first cornerstone for Nothing's full Essential OS |
The consensus, stripped of individual publication vibes: this is a genuinely novel feature that works better than expected, fails in predictable ways, and points toward something that could matter a great deal if Nothing executes on the longer roadmap.
What Is Good — and Genuinely So
I wanted to be cynical about this. The AI hype cycle has earned that cynicism. But there are three things here that I think deserve honest praise.
First, the interaction model is right. Typing "I want a countdown timer for my tea" and getting a countdown timer for your tea is how technology should work. We have spent years navigating someone else's information architecture — someone else's menus, someone else's settings, someone else's idea of what a timer app should look like. The idea that your phone should just do the thing you described is not revolutionary. It is overdue.
Second, the community sharing model is smart. Most people will not want to create widgets from scratch. They will want to browse what others have made and install the good ones. Nothing's Playground community is essentially an app store where the apps are made by users, using AI, and can be forked and customised. That is a different economic model from anything Apple or Google has built.
Third, Nothing is shipping, not teasing. The beta is live. The waitlist is open. Phone (3) owners can try it today. In an industry drowning in concept videos and "coming later this year" promises, there is something refreshing about a company that stands on a stage and says: here, try it now.
What Is Not Good — and Honestly So
Here is the thing, though.
These are widgets, not apps. The distinction is not pedantic. A widget sits on your homescreen and displays information or provides a simple interaction. An app can navigate between screens, store persistent data, access device sensors, communicate with servers, and handle complex workflows. Essential Apps, today, produces widgets. Nothing calls them apps. Every reviewer noticed the gap between the name and the reality.
Phone (3) only. If you own a Nothing Phone (2), a Phone (2a), or any CMF device, Essential Apps does not exist for you yet. Nothing says broader rollout is coming, but has not committed to a timeline. For a company that has sold seven million devices, limiting a flagship feature to your newest — and most expensive — handset is a choice that deserves scrutiny.
Beta means beta. Reviewers reported widgets that looked wrong, prompts that were misinterpreted, and generated code that occasionally did not work at all. That is expected for a beta. It is also expected that some users will try it, encounter the rough edges, and write the whole concept off before it matures. First impressions are difficult to reverse.
The 3/10 repairability problem. This is where my perspective as someone in the repair industry becomes relevant. The Phone (3) scored 3 out of 10 on iFixit's repairability scale. Glued components, hidden screws, what iFixit described as "visually destructive" disassembly. Nothing is building ambitious software for a device that is notably difficult to fix. The Phone (3a) scores slightly better at 4.5/10, and the older Phone (2) managed around 7/10. The trajectory is going the wrong direction.
A company that wants to build phones that "adapt to people" might consider making phones that people can actually maintain.
The Bigger Picture: Essential OS and the Billion-App Bet
Essential Apps is not the destination. It is the first step toward something considerably more ambitious.
Nothing's stated roadmap leads to Essential OS — a full AI-native operating system where the entire interface is generative. Not just homescreen widgets, but the full phone experience shaped around individual users. Carl Pei describes the vision as "one billion apps for one billion people" — the idea that every person should have software tailored precisely to their needs, generated in real time, rather than choosing from a finite catalogue of pre-built options.
The $200 million raised in September 2025 was earmarked specifically for this. Not for phone hardware. Not for earbuds. For the AI platform.
Heise, the respected German tech publication, framed today's launch as "the first cornerstone for Essential OS." That is probably the most accurate description. What launched today is a foundation, not a finished product. The question is whether the foundation is strong enough to build on.
The competitive context is important here. Apple's approach to AI — Apple Intelligence — has been conservative and incremental, focused on making existing tasks slightly easier. Google's Gemini integration is broader but still fundamentally about enhancing the traditional app model. Neither Apple nor Google is asking whether the app model itself needs rethinking.
Nothing is. That takes nerve, especially from a company that is a fraction of the size of either competitor.
"In a market where Apple and Google have spent fifteen years training us to browse someone else's app store, Nothing is asking a genuinely different question: what if you could just make the app yourself?"
What Vibe Coding Means — and Does Not Mean
The term "vibe coding" deserves a moment of scrutiny, because it is going to be everywhere for the next six months and most people using it will not be precise about what it means.
Vibe coding, in the developer community where it originated, means using AI to generate functional code from natural language descriptions. You describe the vibe — the intent, the feeling, the rough shape of what you want — and the AI handles the implementation. It is not a Nothing invention. It is a broader trend in software development that Nothing is applying to consumer smartphones.
What makes Nothing's implementation interesting is who it is aimed at. Vibe coding in developer tools targets people who already understand software. Essential Apps targets people who have never written a line of code and never will. That is a fundamentally different challenge, and one where the prompt design — how the system interprets imprecise, everyday language — matters more than the raw code generation capability.
Early reviews suggest Nothing has done a reasonable job here. The system understood "tea timer that screams," which requires parsing intent (countdown functionality), output specification (audio alert), and personality (the scream, not a gentle chime). That it sometimes fails with more complex or ambiguous requests is unsurprising. Natural language is messy. Human intent is messier.
What This Means for Nothing Phone Owners
If you own a Phone (3): you can join the beta waitlist today. Expect a feature that is fun to experiment with, occasionally useful, and not yet reliable enough to depend on. Treat it as a preview of where Nothing is heading, not a finished tool.
If you own an older Nothing phone: wait. Nothing has signalled that Essential features will expand to other devices, but there is no confirmed timeline. Your phone is not suddenly obsolete — Nothing OS continues to receive updates, and the core Android experience is unchanged.
If you are considering a Nothing phone: Essential Apps is a reason to be interested, not a reason to buy. The Phone (3) starts at £649 in the UK. That is a significant outlay for a beta AI feature. If the rest of the phone's hardware and software appeal to you independently, Essential Apps is a compelling bonus. If you are buying specifically for Essential Apps, wait for it to leave beta.
On Repairs and Longevity
Nothing phones — like all smartphones — eventually need servicing. Screens crack. Batteries degrade. Charging ports wear out. The Phone (3)'s poor repairability score means that when something does go wrong, it takes specialist knowledge and tooling to fix properly. The older Phone (1) and Phone (2) are considerably easier to work on, with well-established parts supply chains through third-party suppliers across Europe.
At celltech, we cover the full Nothing range — from the original Phone (1) through to current models. If your Nothing phone needs attention, whether it is running Essential Apps or not, that is something we can help with.
The Verdict, for Now
Is Essential Apps polished? No. Are the generated widgets occasionally baffling? Absolutely. Is the gap between "Essential Apps" and "Essential Widgets" big enough to matter? Yes, and Nothing should close it quickly.
But the underlying concept — that your phone should adapt to you, not the other way around — is exactly the kind of thinking that has been missing from mobile for years. In a market dominated by two companies that have had fifteen years to get comfortable, a London-based upstart asking uncomfortable questions about the app model itself is, at minimum, interesting. At maximum, it is the beginning of something that makes the last decade of smartphone stagnation look like the warm-up act.
This is either the beginning of something significant, or an expensive proof of concept. Right now, honestly, it is both.
The demo started with a screaming tea timer. Silly. Impractical. Slightly absurd. But someone described what they wanted, and thirty seconds later, it existed. No app store. No download. No compromise.
If Nothing can make that work for the things that actually matter — not just timers that scream, but tools that think — they will have earned every penny of that $1.3 billion valuation.
For now, the tea is still steeping.