A1Base spent its early days handing AI agents a phone number and an email. Then it noticed the harder problem was distribution - and turned itself into a social network where anyone can ship an app in minutes. This is the story of that turn.
There is a sentence you hear in every dorm, every group chat, every startup that never gets built: someone should make an app for this. A1Zap's entire wager is that the someone is you, and that it should take an afternoon rather than a semester. The company - built by A1Base, a small consumer-AI outfit in Y Combinator's Winter 2025 batch - calls itself a social network where people build things for each other. Not photos. Not status updates. Working software.
That framing is newer than it sounds. A1Base did not start as a social network. It started as plumbing. The first pitch was clean and a little dry: an API that gives AI agents a phone number, an email address, and a trusted identity, so they could talk to people over WhatsApp, SMS, iMessage and Slack instead of living inside a chat window. Useful for developers. Invisible to everyone else.
"You should have an AI that's on your side, not just the company's side."
Pasha Rayan, CEO & Co-FounderSomewhere inside the batch, the founders noticed the plumbing was solving the wrong half of the problem. Giving an agent an identity was the easy part. The hard part - the part nobody had cracked - was getting the thing in front of people. So they flipped the company on its axis. Instead of shipping infrastructure and hoping developers found an audience, they built the audience into the product. Distribution stopped being the thing you figure out later. It became the thing itself.
On A1Zap, you describe what you want and the platform assembles it - a chatbot, a small tool, a game, an agent - then hands you a link to drop into a group chat. No code editor. No app-store review. No two-week wait for a human to approve your icon. The early examples are deliberately unglamorous: a Rate My Rent tool for a neighborhood, a matchmaker for a friend group, a campus map, a class-schedule swapper. None of these are venture-scale on their own. All of them are exactly what a student would build in a free afternoon if the friction were gone.
That is the quiet insight underneath the product. Most software never gets made not because the idea is hard, but because the distance between idea and shipped thing is too long. A1Zap is a bet on collapsing that distance to near zero, and then watching what a few thousand bored, clever, hyper-connected students do with it.
A1Zap's first market is college campuses, and the choice is not sentimental. A campus is a rare thing: a dense, captive audience that shares problems, trusts word-of-mouth, and has no patience for bad tools. If a matchmaker or a rent tracker is any good, it spreads across a dorm in hours. If it is bad, it dies just as fast. That is the tightest feedback loop a young social product could ask for, and it is the same terrain the founders know cold from their last company.
The people it serves come in two flavors. On the surface are the makers - students who want to build something for their friends without becoming programmers first. Underneath are developers, who can still reach the original A1Base API to give their own agents a real-world identity and a way to send messages across platforms. One product, two audiences, stacked: a consumer social network sitting on top of the infrastructure the company already knows how to build.
Every builder pays a tax, and it is not the code. It is getting anyone to see the thing. The old app economy charged that tax twice: a gatekeeper who decided what shipped, and a 30 percent cut on the way out. A1Zap's answer is to remove both. Because the app is born inside a social graph, distribution and creation happen in the same motion. You do not build and then go looking for users. You build for the users, in the place they already are.
"The previous generation of messaging apps just weren't built for what was coming in the future."
Pasha RayanThe pivot is the most instructive part of the whole story, and it is worth walking through plainly.
Notice what did not happen. The team did not scrap the old work. The messaging rails, the agent identity, the multi-channel delivery - all of it still runs underneath A1Zap. The pivot was not a demolition. It was a change in which floor of the building the customer walks into.
Pasha Rayan and Pennie Li met the way a lot of founders do - at university, in this case UNSW in Sydney - before ending up in San Francisco. Rayan is not new to this. His last company, Forage (which began life as Inside Sherpa), turned virtual job simulations into a platform that reached roughly ten million users and worked with more than five thousand university partners before it was acquired in 2024. He has already built one thing that touched a lot of students. A1Zap is, in a sense, him aiming at the same audience with a very different tool.
"I didn't want to be a consultant - I wanted to build things."
Pasha RayanThe company describes what it is chasing as "warm AI" - technology, in its words, that makes being human feel more human. It is a deliberate word choice in an industry that mostly advertises how well its agents replace people. A1Base's framing is the opposite: the agent is a collaborator inside a conversation between humans, not a substitute for one. Whether users end up caring about that distinction is exactly the kind of question an early-stage bet exists to answer.
There is a growing shelf of tools that let you conjure software from a prompt - the Replits and Lovables and Bolts of the world - and a separate shelf of places to talk to AI. A1Zap is trying to be a third thing: not a builder you visit alone, and not a chat you consume, but a feed whose core loop is making rather than scrolling. Most social products reward you for watching. This one is designed to reward you for shipping. That is a small design decision with an enormous behavioral consequence, and it is the whole gamble.
It is early. The team is small, the funding is a seed round, and the map beyond campus - toward cities and local communities - is still a sketch. But the shape of the bet is clear and unusually legible: strip out the friction, fold distribution into the network, start where feedback is fastest, and see whether "someone should build this" quietly turns into "I just did."