In late 2009 a teenager in Moscow built a website that connected two random webcams. Within months it had gone from a handful of users to a genuine global phenomenon, covered by every major newspaper and parodied on television.
Then it curdled — and the way it curdled is the single most instructive story in this whole category.
The rise
Chatroulette’s appeal was the same as Omegle’s but sharper: pure serendipity, with faces. You clicked, you saw a person somewhere in the world, and either of you could move on instantly. Nothing else online offered that.
Growth was extraordinary because the format is inherently viral. Every session produces a story worth telling, and every story sends more people to try it. Musicians performed on it. Comedians built sketches from it. It became a cultural reference point within about a quarter.
The problem that arrived with the users
The failure was rapid and it was structural: a large and highly visible share of users were men exposing themselves on camera.
This was not a small minority behaviour that could be moderated at the edges. It was common enough that it defined the experience. Anyone recommending the site had to add a warning, and eventually the warning became the reputation.
The mechanics are worth stating plainly, because they generalise. Anonymous live video with no account and no consequence selects for exhibitionism. There is no cost to being removed, because there is nothing to remove — a ban is a page refresh. And moderation cannot keep up, because reviewing live video at volume in real time is not a solvable problem for a small team.
What they tried
Chatroulette did not ignore it. Over the years it tried automated image detection to flag nudity, user reporting and voting systems, temporary bans, and eventually gating parts of the service behind sign-in and verification.
Some of it helped. None of it restored the original experience, because by then the reputation had set. The people who left over the content did not come back to check whether it had improved. That is the part founders underestimate: you get one reputation in this category, and cleaning up afterwards is far harder than being clean from the start.
Why Omegle lasted longer
Omegle survived considerably longer, partly because it kept a text mode alongside video, which gave people a way to use it that did not involve a camera lottery.
But it ended up in the same place for the same underlying reason, just on a longer timeline and with worse consequences — the litigation that finally closed it in 2023 concerned harm to a minor on the video product. We covered that story in why Omegle shut down.
Two of the biggest names in random chat, both undone by the same feature.
Is Chatroulette still running?
Yes, in a considerably smaller and more controlled form. It never fully disappeared — it added verification and filtering, shed most of its mainstream audience, and settled into a niche.
That trajectory is worth noticing on its own. The site did not die; its moment died. Once a platform in this category becomes known for one thing, that reputation is close to permanent, and no amount of subsequent moderation work recovers the audience that left.
Why the growth was never sustainable
There is a second lesson underneath the content problem, and it applies to every random chat platform since.
Chatroulette’s explosive growth was driven by novelty and press coverage, not by retention. People tried it because it was the thing everyone was talking about. Very few built a habit around it, because the format has no reason to bring you back tomorrow — there is no feed, no friends list, no notification, nothing waiting for you.
That means the category is unusually dependent on constant new attention, and unusually fragile when the attention turns negative. A platform with strong retention can survive a bad news cycle. One running purely on novelty cannot.
It is also why liquidity is such a persistent problem in this space — covered in more detail in how matching actually works.
The lesson worth taking
The category’s repeated failure is not a coincidence and not bad luck. It follows from a specific combination: anonymous, live, visual, unaccountable. Remove any one of those and the failure mode weakens dramatically.
Remove visual and abuse becomes filterable text. Remove anonymous to the platform — keeping anonymity between users — and bans start to mean something. That is the design most surviving platforms have converged on, and it is the reasoning behind choosing text over video.
The insight both sites were built on was genuinely good. Meeting a random person from anywhere is a real thing people want. What neither survived was the delivery mechanism.
If you want the current landscape, we keep a ranked list of alternatives — and the first thing worth checking on any of them is which of those four properties they kept.