AutoMod

Automatic content moderation without manual intervention

AutoMod monitors every message on your server and responds automatically when a filter is triggered. For each filter you can define what should happen: delete, warn, mute, or all of the above. Trust roles and channel exceptions ensure that Moderators and certain areas are never filtered.

Setup

  1. Enable AutoMod in the dashboard
  2. Enable and configure the desired filters
  3. Select an action per filter (off / delete / warn / delete_and_warn)
  4. Optional: add bypass channels and trust roles
  5. Save

Available filters

Each filter can be enabled independently. Per filter you choose an action and optionally exception channels.

Profanity (badWords)

Checks messages against a self-maintained word list. Match mode: partial (word appears anywhere) or exact (whole word). Case-insensitive by default. Up to 500 entries, each max. 100 characters.

Repeated text (repeatedText)

Detects when the same character appears more than N times in a row (default: 10). Useful against "aaaaaaaaaa"-style spam.

Server invites (invites)

Blocks Discord invite links. The allowOwnServer option (default: true) permits links to your own server. You can maintain a whitelist of allowed server IDs.

External links (externalLinks)

Filters all external URLs. With a whitelist you can add allowed domains (e.g. "twitch.tv"). blockAll: true blocks all URLs without exception.

Caps (caps)

Triggers when more than X% of letters are uppercase (default: 70%) and the message is at least N characters long (default: 10). Guards against SHOUTING.

Too many emojis (emojis)

Limits the number of emojis per message (default: 5). Counts both Unicode emojis and custom emojis.

Spoiler abuse (spoilers)

Limits the number of ||spoiler|| blocks per message (default: 3). Useful against spoiler spam.

Mass mentions (mentions)

Triggers when too many users are mentioned in a single message (default: 5). @everyone and @here count double (configurable).

Zalgo text (zalgo)

Detects combining Unicode characters used for visual chaos text. Threshold: percentage of characters with diacritic stacking (default: 30%).

Anti-spam (antiSpam)

Triggers when a user sends N messages within X seconds (default: 5 messages in 3 seconds). Prevents message flooding.

Actions

ActionWhat happens
offFilter inactive — no action
deleteMessage is silently deleted
warnUser receives a warning, message stays
delete_and_warnMessage deleted + warning issued

Trust roles and bypass channels

All members holding a configured Moderator role are automatically exempt from all filters. Additionally, you can mark specific channels as bypass channels per filter — in those channels the filter is ignored for all users.

Add bot spam channels as a bypass channel in the link filter so command responses containing URLs are not filtered.

Evasion attempts

TeraOne detects common typographic evasion attempts for the profanity filter in partial-match mode. The exact detection mechanisms are deliberately not documented. If you notice a user actively bypassing the filter, report it to TeraOne support.

Frequently asked questions

Is AutoMod applied in voice channels?
No, AutoMod only monitors text messages. Voice activity is not filtered.
Do messages in threads count towards anti-spam?
Yes. Messages in threads and forums are treated the same as regular channel messages.
Can I use AutoMod alongside Discord's built-in AutoMod at the same time?
Yes, both run independently. If both are active and both trigger, the action may happen twice. Recommendation: use either TeraOne AutoMod or Discord's built-in AutoMod for the same filter type.
How long are AutoMod incidents stored?
AutoMod actions that result in warnings are stored permanently in the user's mod history. Silent deletions (delete only) are not logged permanently.
AutoMod · TeraOne Docs · TeraOne