If you came online between roughly 1997 and 2003, there’s a reasonable chance your earliest friendships with people you’d never met in person were tracked by a six- or seven-digit number. That number was your ICQ UIN — your Universal Internet Number — and it was, for a brief and genuinely strange window in internet history, the way millions of people understood their online identity.

ICQ was the first mass-market real-time chat service that worked across the open internet rather than within a walled-garden service like AOL or CompuServe. It came out of Israel in November 1996, built by four young programmers at a company called Mirabilis. Within eighteen months it had over twelve million registered users. By the time AOL bought Mirabilis in June 1998 for $407 million in cash plus performance bonuses, ICQ had grown to roughly twenty-six million users — a stunning trajectory for the time.

This is the story of how that happened, and how the service that defined a generation of online communication ended up as a footnote.

Where it came from

The four Mirabilis founders — Yair Goldfinger, Arik Vardi, Sefi Vigiser, and Amnon Amir — built the first version of ICQ in a few months in late 1996. The product addressed something obvious that nobody else had solved cleanly: how do you find your friends online and chat with them in real time, without both of you having to be in the same chat room or on the same proprietary service?

The solution was the contact list. You’d add people by their UIN, the system would tell you when they came online, and you could start a conversation. The interface was minimal. The famous notification sound — that particular “uh-oh” — became one of the most recognizable audio cues in computing history, comparable to the Windows startup chime or the AOL “you’ve got mail.”

The viral mechanic was simple. When you installed ICQ, the first thing you wanted was friends to chat with. So you sent your UIN to people you knew. They installed it to talk to you. They sent their UIN to others. The growth curve was exponential, and the founders documented in early interviews that they were genuinely surprised by how quickly it took off.

The AOL acquisition and what changed

In 1998, AOL was the dominant force in consumer internet access in the US, and AOL Instant Messenger (AIM) was its own messaging product. The Mirabilis acquisition was framed as a strategic move to acquire ICQ’s international user base and prevent a competitor from doing the same.

The reality of what happened post-acquisition is more complicated. AOL ran ICQ and AIM as separate products with separate codebases. Interoperability between the two — which would have made obvious product sense — never quite happened, partly because of internal politics and partly because of the technical headaches of merging two protocols designed by different teams.

ICQ continued to grow internationally, particularly in Russia, Eastern Europe, Germany, and parts of Latin America. The product accumulated features at a steady pace through the early 2000s — file transfer, voice chat, integrated games, multi-user chat. But the core experience didn’t evolve fast enough to compete with what was coming.

What killed it

A few things, layered together.

MSN Messenger launched in 1999 and benefited from being bundled with Windows. The friction of installing ICQ separately mattered when there was already a chat client sitting on every desktop.

Then came the social network and the mobile shift. MySpace, Facebook, and later WhatsApp restructured how people thought about online identity. The numeric UIN — once a kind of cool internet handle — started to feel awkward next to a real name and photo. And ICQ’s mobile transition, when it finally happened, was clumsy.

The product passed through several owners. AOL sold ICQ to the Russian investment group Digital Sky Technologies in 2010 for $187.5 million — less than half what it had paid in 1998. DST eventually became Mail.ru Group, which kept ICQ alive primarily as a service for its remaining Russian-speaking user base.

In June 2024, ICQ was officially shut down. The announcement came quietly. The servers went dark. A nearly twenty-eight-year run ended.

What ICQ actually contributed

It’s tempting to remember ICQ as a curiosity, but the design vocabulary it established has had remarkable staying power. The persistent contact list is now the foundational metaphor of every chat application in existence. The presence indicators — online, away, busy, invisible — were largely standardized by ICQ. The notification sound as identity, the away message as performance art, the careful management of who could see your status: all of these patterns moved from ICQ into AIM, then MSN, then Skype, then WhatsApp and Slack and Discord.

For internet historians, the more interesting question is what ICQ tells us about the dynamics of platform succession. The service had massive user numbers, real product innovation, and a strong international presence. None of it was enough to survive a series of strategic mistakes by absentee owners and a structural shift in how people used the internet.

The Internet Archive’s Wayback Machine preserves snapshots of ICQ.com going back to 1997, and they’re worth a browse if you want to see how the visual design and feature set evolved across nearly three decades. The 1998 homepage still has that distinct early-web optimism — small graphics, centered text, the assumption that the next million users were just an installer download away.

For a generation, ICQ was the front door to internet sociability. That it ended in silence in 2024 is, in its own way, fitting.