Type a sentence. Get a real app - design, database, login, and a checkout that takes money. The catch is there is no catch, only a bill at the end of the month.
Somewhere right now, a salon owner is building a booking system instead of learning to code. She will not thank a single engineer, because there isn't one.
It is a Tuesday, and roughly 700,000 people who cannot write a line of code have an app builder open. One of them is a real-estate agent in Los Angeles charging clients $100 a month for a property portal she described out loud. Another is a finance worker in Japan who has earned about $34,000 from a suite of tools that did not exist last quarter. They are using Anything, a San Francisco company that turns plain language into software that actually ships.
The pitch sounds like every other AI demo until you read the fine print: Anything does not stop at the pretty mockup. It hands you a live app, with a backend, user accounts, payments, and AI baked in. The prototype is the easy part, and everyone can do it. Anything is betting the whole company on the boring part - the launch.
"Most AI coding tools make great prototypes, but fail on real launches. We've solved the fundamental problem."
- Dhruv Amin, Co-founder & CEOThe hour when an idea meets an error message and one of them does not survive.
There is a particular kind of despair familiar to anyone who has tried to build software without knowing how. You get 90% of the way there. The thing almost works. Then a bug appears at 2 a.m., the error message is in a language you do not speak, and the dream quietly dies in a browser tab. The industry has a polite name for this gap. Anything calls it the doom loop, which is more honest.
For years, "no-code" tools promised to close that gap and mostly delivered wireframes - beautiful, lifeless, and incapable of charging a customer. The demo always worked. The launch rarely did. The problem was never imagination. It was the last mile: databases, authentication, payments, deployment, and the hundred small failures between an idea and a working business.
"We're not heading toward a world where everyone can code more easily, but one where almost no one needs to code."
- Marcus Lowe, Co-founderThey met at Google and were best men at each other's weddings, which is the only acceptable reason to start a company together.
Dhruv Amin was the first product manager on YouTube TV. Marcus Lowe came out of Google Maps. They had watched plenty of impressive technology fail to reach an ordinary person's hands, and they made a specific wager: the future is not everyone learning to code, it is almost nobody needing to. That is a bigger claim than it sounds. It assumes the bottleneck was never the typing - it was everything that happens after.
So they built the unglamorous parts first. The company started in 2021 as a code-generation system for app primitives, did about $1M in revenue by 2022, and then ChatGPT arrived and rewrote the rules. They pivoted hard into AI agents, launched under the name create.xyz, and eventually rebranded to Anything - a name with the confidence of a company that means it.
First product manager on YouTube TV before betting that the hard part of software was the launch, not the prototype.
Came from Google Maps. Argues the endgame is a world where building software no longer requires writing code.
The boring infrastructure, wired in so quietly that the user never learns it exists.
Anything takes a prompt and returns a working application for web, iOS, and Android. The design is generated. The database is provisioned. Logins work. Payments work. And because no modern app is complete without it, you can drop in GPT-5, Claude, Gemini, DALL-E 3, or Stable Diffusion with a single sentence and no API keys - more than 40 models, handled behind the curtain.
Describe an app in plain English; get a live one. Frontend, backend, database, auth, and payments generated and deployed together.
An AI software engineer that opens your app in a real browser, finds the bug, rewrites the code, and keeps going until it works.
GPT-5, GPT-4o, Claude, Gemini, DALL-E 3, Stable Diffusion - one prompt, no keys, no plumbing.
Multi-platform deployment plus built-in monetization, so the gap between "I made a thing" and "I have a business" closes.
What people actually build with it:
"Anything handles the code, design, database, payments, and even AI integrations - so you can focus on your idea and launch fast."
- From the Anything platformFive years, three names, and one stubborn idea about the launch.
An initial code-generation system for app primitives. The unglamorous foundation gets poured first.
Hits $1M revenue; backed by Uncork Capital. Then ChatGPT lands and changes the threshold of what's possible.
Reaches $2M revenue and pivots into AI agents. Bessemer comes aboard in 2024.
Launches as create.xyz, adds the pieces, and rebrands to Anything.
Public launch on August 7. Reaches roughly $2M ARR within two weeks.
Closes an $11M Series A at a $100M valuation led by Footwork, and ships Anything Max.
Two weeks is a short time to build a business and a long time to fake one. They did the first.
Growth that fast usually comes with an asterisk. Here the asterisk is smaller than the headline. Anything reached about $2M in annualized recurring revenue within two weeks of its August 2025 launch, crossed 700,000 registered users, and raised $11M at a $100M valuation led by Footwork, with M13, Bessemer, Uncork, and Shopify CEO Tobi Lutke along for the ride. Total capital raised now sits near $19.5M.
"Anything hit $2M ARR in 2 weeks, then raised $11M at a $100M valuation."
- TechBuzz.ai, on the launchThe plan is not to make coding easier. The plan is to make it optional.
Strip away the funding noise and the mission is plain: let anyone turn an idea into a working, paying business without hiring an engineer or becoming one. It is a redistribution of who gets to build. For decades, software was a craft guild with a high wall around it. Anything is trying to take the wall down, one prompt at a time, and it is doing it by handling the parts nobody enjoys.
Whether "almost no one needs to code" turns out to be prophecy or marketing is the open question. Skeptics will note that production software is hard for reasons that do not vanish because an agent is fast. Believers will point at 700,000 people who already disagree. Both can be right for a while.
"Type what you want. Get a live business. That is the entire bet, and it is a big one."
- The Anything thesis, paraphrasedThe agent in Los Angeles is still working. So is the one in Japan. Neither has met the other, and both have shipped.
If the next decade of software belongs to people describing what they want instead of typing how to build it, the companies that win will be the ones that closed the last mile - the database, the login, the payment, the 2 a.m. bug. That is precisely the unglamorous ground Anything chose to stand on.
So return to that Tuesday. The real-estate agent in Los Angeles still cannot code. The difference is that it no longer matters. Her portal is live, the rent on her software is paid, and the doom loop that would have killed it at 2 a.m. is now a bug an agent fixes while she sleeps. The wall around software just got a door. Anything is selling the key.