Safety Center

Before a chat starts

• Adults only. Everyone confirms they are 18 or older before their first chat.
• Real people. A "verify you're human" check, confirmed by our server with Cloudflare, keeps bots and scripts out.
• No VPNs or proxies. They are how banned people try to come back.
• A ban follows the person, not just the browser. Clearing your browser data does not undo a ban, a block or a report: the device is recognised by a one-way hash and its network, and the identity is carried over. Separately, a new browser from a network that a banned person recently used starts in the low-trust group and has to pass the human check far more often, until it has earned trust. Nobody else on a shared network is banned because of it.

During a video chat

• Off and blurred until you both tap Show. Every video chat starts with your camera, microphone and the stranger's speaker off, and the pictures blurred. Only when both of you tap Show do they all come on together; skipping or exiting turns everything off again. The blur is applied in your own browser before your video is sent, so the other person never receives a clear picture before you agree.
• Hide your face. Blur your face, or wear a cartoon avatar that follows your head. If your face can't be found, your whole picture is hidden — never shown clear.
• Nudity check. Your browser checks the stranger's video and blurs it if it looks explicit. Our server also checks occasional stills of what you receive with an AI safety model: a share of all chats, and every chat with someone who has been reported or is in the low-trust group. A flagged still blurs the video for you at once, with a red warning, and is kept for moderators; everything else is thrown away immediately, and a chat that has looked fine for a minute or so is no longer checked. The first detection warns the person on their own screen. A second one, from a different viewer, is a ban.
• Data saver. On a weak connection video pauses and the call carries on as voice.
• Encrypted. Video and voice calls use WebRTC, which encrypts media end to end between the two browsers. Text goes through our server over an encrypted connection and is not stored — unless it is part of a report.

In every chat

• Explicit messages never arrive. Every line is checked before it is delivered — in chats, spectator mode, voice-room text, live captions and the interests you type. It sees through swapped, masked, stretched and spaced-out letters, and works in English, Hinglish and Devanagari. A blocked line is not shown to anyone, the sender is banned. Anything that involves a minor is a permanent ban and goes straight to a moderator.
• Spam is a ban. Copy-pasted promotion, links and phone numbers sent to several strangers get the sender banned, not just limited.
• GIFs come from GIPHY only. Pictures cannot be uploaded and sent, so a nude image cannot be slipped past the text filters as a GIF.
• No links or phone numbers. They are removed before delivery — that is how scams and grooming move a chat off the site. Abusive words, including romanised Hindi, are masked.
• Automatic bans, no mutes. There is no muting: explicit or abusive messages, flooding the chat (more than eight messages in ten seconds), spam, and nudity on camera each lead to a ban. Swearing that the filter catches is blocked, and three filtered messages in a day is a seven-day ban.
• Trusted badge. People who chat regularly without being reported earn it, and can choose to meet only other trusted people.
• Repeat offenders are kept apart. People reported by several others are only matched with each other and are kept out of voice rooms.
• Block, everywhere. Someone you block is never put with you again anywhere on Anyeria — not as a random match, not in a voice room, and not in spectator mode, whichever of the three roles either of you is in. It follows the browser rather than the network address, so blocking one person on a shared mobile network does not make everybody else on it vanish.

In audio chat

• Voice only. There is no camera in audio chat, so nothing can be seen.
• You start muted. Your microphone is off for every call until you unmute it, and goes back to muted when the call ends or you skip.
• Same protections. Report and block are on the screen, text is checked like everywhere else, and the call is encrypted between the two browsers.

In spectator mode

• Report and block anyone. Both buttons sit in the corners of the conversation, and both ask which of the other two you mean — the stranger, or the person watching. The two talking can act on the spectator just as the spectator can act on them.
• The messages come with the report. The last lines of the three-way conversation are attached by our server, exactly as in any other chat.
• The watcher cannot speak. That is enforced on our server, not just hidden in the page. They can add an emoji to a line, which is the whole of what they can say.
• Everyone knows they are being watched. The two talking are told there is a spectator and where they are connecting from.

In voice rooms

• Vote someone out. A majority of the room can remove someone; one person alone never can.
• You join muted. Your microphone stays off until you tap it.
• Mute anyone, just for you, without involving the rest of the room.
• Report or block from the room. A report carries the room's recent messages between you and the person you are reporting.
• Invitation links. A room link lets a friend join the room you are in. It does not bypass anything: everyone who follows one still confirms they are 18 or older and still passes the human check, and anybody you have blocked is never placed in that room.

Choosing who you meet

The paid Country, Gender & Age Filter is a matching preference, not a safety feature, and we are careful not to let it read as one. Saying who you want to meet does not verify that anybody is who they say they are — gender and age are stated by each person about themselves and are not checked. Treat everyone you meet as a stranger, because they are. If somebody's stated age looks false, report them.

Stating your own gender and age is free, optional, held only for the length of a connection, and never shown to the people you are matched with. See the Privacy Policy.

When you report someone

1. Pick what happened, and add a few words if you like. Reports are anonymous — the person is not told.
2. The last messages of the chat are attached automatically, straight from our server, so nobody can fake them. In a video chat you can also attach a still of their video; only moderators see it, and it is deleted after 30 days.
3. Reports about minors and nudity are reviewed first. A person reported as under 18 by a trusted user, or by people on two different networks, is suspended immediately while a moderator looks.
4. When enough different people — on different networks — report the same person, they are banned automatically for a short time. A single group of friends cannot trigger this.

What happens to people who break the rules

Penalties climb with each offence, and are forgotten after 90 days without another:

warning → 1 hour → 24 hours → 7 days → permanent ban

Most automatic rules go straight to a ban and use the ladder from a 24-hour minimum (flooding, spam behaviour). Sexually explicit or abusive messages and nudity (confirmed by a second viewer) are a seven-day ban the first time and permanent for a repeat. Repeated filtered language is a seven-day ban. Anything sexual involving a minor is a permanent ban or an immediate suspension while a moderator looks.

The person is always told why. Anyone banned can appeal once from the page they see, and a moderator reads every appeal. An appeal that is accepted lifts the ban and removes the strike. Automatic decisions are checked by people too: a moderator can mark any of them as a false positive, which lifts the ban and takes the strike back.

Talk to a person

Our Grievance Officer is Satyam Kumar Verman — [email protected]. Complaints are acknowledged within 24 hours and resolved within 15 days; content that shows someone's private areas or impersonates them is removed within 24 hours. More on the Contact page.

Child sexual abuse material is reported to the authorities:

• National Cyber Crime Reporting Portal (India) — or call 1930
• NCMEC CyberTipline — international

The full rules are in the Community Guidelines.