Content Moderation Policy
Last updated: 11 November 2026
This policy explains what players can create on papergames.io, what is not allowed, how content is checked automatically and by people, and what happens when someone breaks the rules. It is written for players and for our advertising and demand partners. It sits alongside our Terms and Conditions and our Privacy Policy.
1. Where players can create content
Players can create content in these places, and only these:
- Display names. Every account has a display name. It is shown on profile pages, which search engines can index, on leaderboards, and in tournaments and game rooms.
- Private inbox chat. Two players can message each other only once they are friends: one sends a friend request and the other accepts it. Inbox messages are stored so the conversation is still there when the players come back.
- In-game chat. Only the players in a game see its chat. Spectators do not. In-game messages are never stored.
- Tournament names and descriptions. A registered member who organises a tournament gives it a name and, optionally, a description. Both are checked before the tournament is saved.
- Report comments. A player who reports someone under "Other" can add a comment. Only the moderation team sees it.
There are no uploads of any kind: players cannot share images, video, audio or files, and a link in a message is shown as plain text, never as a preview or embed. Avatars come from a fixed catalogue owned by papergames.io.
Chat needs an account, either a guest account or a registered one. A visitor without an account can watch games but cannot send any message.
2. What is not allowed
Display names
A display name must not convey any of the following, in any language. This includes names that disguise a word with lookalike characters, numbers, spacing or reversed spelling.
- Slurs or hate speech targeting race, ethnicity, religion, nationality, gender, sexual orientation or disability.
- Sexual or pornographic content, including references to sexual acts or genitalia.
- Harassment, threats, violence, self-harm, or promotion of terrorism or extremist groups.
- Illegal drugs, drug dealing, or grooming and predatory intent.
- Nazi, fascist or other extremist references, including coded ones such as 88 or 1488.
- Impersonation of papergames.io staff, moderators, administrators, support or the platform itself.
- Personal contact details, such as phone numbers or email addresses, or invitations to move to another platform.
Mild trash talk, harmless nicknames, and references to games, memes, brands and real people are allowed, as are words in other languages that are not offensive.
Messages
Inbox and in-game messages must not contain:
- Harassment, including threats against a person.
- Hate, including threats against a group.
- Self-harm content, including statements of intent and instructions.
- Sexual content, and any sexual content involving minors.
- Violence, including graphic violence.
- Advice or instructions for illegal activity, including violent wrongdoing.
Spamming, cheating and other unfair play are not allowed either, and players can report them.
3. Automatic checks
Display names
- A fixed list of banned words is applied when an account is created and on every name change. Any listed word is masked out of the name.
- An AI classifier then reviews the name of every new account about 15 seconds after signup, whatever the signup method, and reviews every name change as soon as it is saved. It rejects a name only when it is more than 85% confident that the name breaks the rules above.
- A rejected name is replaced at once with a generated one, and the moderation team is alerted by email. The first time this happens to a player, they also receive a message in their papergames.io inbox explaining why.
Inbox messages
- Every inbox message is checked by an automated classifier, OpenAI's moderation model, before it is delivered.
- A flagged message is hidden from the recipient and sends them no notification. The sender sees it marked as moderated, and it is reported to the moderation team automatically.
In-game chat
- Every in-game message is checked by the same classifier while it is being delivered, so games are not slowed down. A flagged message is removed from both players' screens within moments and replaced with "Message removed by moderation".
- Inbox and in-game messages are limited to 2,000 characters.
When the classifier is unavailable
If the classifier cannot be reached, messages are still delivered and names are kept without the AI review. They remain open to player reports.
Tournament names and descriptions
- Every time an organiser creates or edits a tournament, its name and description are checked before anything is saved.
- The name is checked against the fixed list of banned words first. A listed word anywhere in the name refuses the save. Descriptions are not checked against the list.
- The same AI classifier that checks inbox messages, OpenAI's moderation model, then reviews the name and the description. If it flags either one, the save is refused.
- When a save is refused, nothing is published, and the organiser is told that the name or description breaks our content rules and asked to reword it.
- If the classifier cannot be reached, the tournament is saved once its name has passed the banned-word check, and the description is saved without review.
- Tournament pages are not indexed by search engines, and players can report the organiser.
4. Reporting
Any player with an account can report another player from the menu on that player's profile panel, in the friends list, or in an inbox conversation. The report form offers these categories:
- Abuse: Verbal abuse / Cursing / Trolling, Racism, Violence / Threats, Sexual Harassment.
- Fair play: Cheating, Stalling / Quitting Games, Sandbagging (Rating Manipulation).
- Other: Offensive Username, Spamming, Other (with an optional comment).
The player can also block the other player in the same step. Blocking closes any inbox conversation between them and stops friend requests from the blocked player.
- A report of racism, sexual harassment or an offensive username goes to a human moderator on the first report.
- Reports in other categories go to a moderator once two different players have reported the same person, and again at five and at ten.
- Repeated reports of the same person by one player count once.
5. Human review and response times
- A human moderator reviews escalated reports within 4 hours.
- Automatic actions, such as replacing a name or hiding and removing a message, take effect as soon as the classifier returns its verdict, without waiting for a person.
- To assess a report, moderators can read the inbox conversation it concerns.
- Each month we review report volumes and how accurately the automated classifiers perform. Every change to the display-name classifier is tested against a set of names labelled by people before it ships.
6. Enforcement
- Name replacement. A name that breaks the rules is replaced with a generated one.
- Message hiding and removal. Flagged inbox messages are hidden from the recipient, and flagged in-game messages are removed from the game.
- Formal warnings. A moderator can give a player a formal warning in any report category. A warning about a display name also replaces the name.
- Automatic ban on a second warning. If a player already has a warning of the same type on record, a new warning from a moderator bans the account automatically. Each name the classifier rejects counts as an offensive username warning.
- What a ban does. A banned account is removed from tournaments and left out of the rankings, is matched only against computer opponents, cannot challenge friends to a game, and cannot send friend requests.
- Moderator bans. A moderator can also ban an account directly. That ban also applies to every other account that shares its IP address. When the player's device is known, the ban records it, and friend requests sent from that device are then ignored, whichever account sends them.
7. For advertising and demand partners
- The only player-created content that can appear on public pages that search engines index is display names, and those are checked as described above.
- Chat is private. Inbox chat is only between friends, and in-game chat only reaches the players in that game. Neither is ever shown on public pages or to search engines.
- Tournament names written by organisers are checked against the banned-word list and the classifier before they are saved, and descriptions against the classifier. Both appear only on pages that carry a noindex tag.
- There is no user-uploaded media.
- The companies that process personal data for papergames.io, including our advertising provider and its partners, are listed on our Sub-processors and data locations page.
Partners with questions about this policy can write to [email protected].
8. Appeals and contact
Players can contest a moderation decision by writing to [email protected]. Please include your display name and describe what happened.
For any other question about this policy, contact us at [email protected]. papergames.io is operated by Poly Bytes Sàrl, Chemin de la Planche-aux-Oies 4d, CH-1000 Lausanne, Switzerland.