feature

Moderation

Configure filters, link protection, warnings, and automatic punishment rules.

Last updated September 7, 2026

What moderation does

Moderation controls automatic checks for words, links, invites, bypass attempts, nicknames, and other member activity. Each enabled control can warn or apply the selected punishment.

Configure moderation

1
Open Advanced Moderation > Moderation.
2
Enable the protections you want to use.
3
Choose a punishment and timeout duration where the setting offers one.
4
Save, then review Action Logs after the feature begins handling messages.

Key settings

Option	What it does
Moderation module	Master switch for the moderation runtime.
Anti-swear level	Sets Light, Standard, or Strict matching.
Language coverage	Free includes English, Spanish, German, and Russian; Premium unlocks every available language pack.
Custom filters	Adds one word, phrase, username, or link per line.
Allow regex rules	Enables advanced expressions; each is capped at 256 characters and unsafe nested quantifiers are rejected.
Invite link removal	Detects standard and obfuscated Discord invites and applies its selected punishment.
Block every untrusted external link	Extends invite filtering to untrusted external links.
Phishing protection	Uses separate Suspicious content and Confirmed malicious content actions.
Suspicious / malicious score	Sets the two confidence thresholds; the malicious threshold controls severe punishment.
Redirect, reputation, and attachment checks	Chooses whether link redirect chains, domain reputation, and attachment metadata contribute to protection.
Allow ordinary safe media	Avoids treating ordinary safe media as suspicious.
Nickname filtering	Applies enabled language and custom filters with Sanitize, Reset, Warn, or Log behavior.
Anti-hoist	Handles leading punctuation, invisible Unicode, symbols, and separators.
Bypass detection	Off matches plain text, Balanced catches spacing/symbol splits, and Aggressive also catches look-alikes/repeats.
Punishment / timeout	Sets the default action and a timeout from 1 to 40,320 minutes.
Trusted scope	Exempts selected roles, channels, categories, users, domains, and invite codes; extra suspicious domains can be added.
Custom warning message	Supports {badword}, {user}, {server}, {reason}, and {channel}; maximum 1,000 characters.
Repost a censored copy	Premium; briefly uses a managed webhook, never reposts attachments or mentions, and requires Manage Webhooks.
Export / Import	Moves moderation configuration as a local file; imported changes are not active until saved.
AI moderation / image filtering	Retained source settings are not surfaced as configurable controls on the current page and are not promised by this guide.

Start cautiously

Key settings

Option	What it does
Moderation module	Master switch for the moderation runtime.
Anti-swear level	Sets Light, Standard, or Strict matching.
Language coverage	Free includes English, Spanish, German, and Russian; Premium unlocks every available language pack.
Custom filters	Adds one word, phrase, username, or link per line.
Allow regex rules	Enables advanced expressions; each is capped at 256 characters and unsafe nested quantifiers are rejected.
Invite link removal	Detects standard and obfuscated Discord invites and applies its selected punishment.
Block every untrusted external link	Extends invite filtering to untrusted external links.
Phishing protection	Uses separate Suspicious content and Confirmed malicious content actions.
Suspicious / malicious score	Sets the two confidence thresholds; the malicious threshold controls severe punishment.
Redirect, reputation, and attachment checks	Chooses whether link redirect chains, domain reputation, and attachment metadata contribute to protection.
Allow ordinary safe media	Avoids treating ordinary safe media as suspicious.
Nickname filtering	Applies enabled language and custom filters with Sanitize, Reset, Warn, or Log behavior.
Anti-hoist	Handles leading punctuation, invisible Unicode, symbols, and separators.
Bypass detection	Off matches plain text, Balanced catches spacing/symbol splits, and Aggressive also catches look-alikes/repeats.
Punishment / timeout	Sets the default action and a timeout from 1 to 40,320 minutes.
Trusted scope	Exempts selected roles, channels, categories, users, domains, and invite codes; extra suspicious domains can be added.
Custom warning message	Supports {badword}, {user}, {server}, {reason}, and {channel}; maximum 1,000 characters.
Repost a censored copy	Premium; briefly uses a managed webhook, never reposts attachments or mentions, and requires Manage Webhooks.
Export / Import	Moves moderation configuration as a local file; imported changes are not active until saved.
AI moderation / image filtering	Retained source settings are not surfaced as configurable controls on the current page and are not promised by this guide.