NewGuardy AI catches scams and spam and sets up your server. Included on every plan, even Free.AI scam and spam filter, free on every plan.

AutoMod · free on every plan

Rules that act before harm spreads.

Guardy Bot checks every message against rules your whole team can read: what each rule matches, who is exempt, and what happens next. Repeat offenders climb a strike ladder you set.

Protected in one click
Five ready protections against scams and spam
Repeats escalate
Strikes climb a ladder you set
AI for what rules miss
Optional AI Shield, on every plan
Safety is free
Every rule and protection on every plan

01Words and patterns

Match precisely

Catch a word, a whole list, or a pattern, and skip the people and places you trust.

  • Keywords and word lists of up to 5,000 entries, matched as whole words or anywhere, with common tricks to dodge the filter caught too.
  • Patterns (regex) are checked when you save. A pattern that runs too slowly is paused on its own, so the rest of moderation keeps running.
  • Each rule can skip roles and channels. Owners, admins and bots are always skipped.

Example rule settings on the dashboard.

02Spam and floods

Stop floods a word list would miss

Set how many is too many, and over how many seconds.

  • Check messages, duplicates, mentions, emoji or reactions in a time window, plus capital letters, spoilers and stacked accents.
  • Turn on all five ready protections in one click: scam and nitro bait, mass mentions, message floods, shouting in caps and emoji spam.
  • Each rule can delete, warn, time out, kick or ban. The member gets a private DM with the reason.

Two of the five ready protections, as they read on the dashboard.

04AI Shield

AI Shield for what rules miss

Optional on every plan. Your rules run first; AI Shield checks the messages they find suspicious but cannot decide on.

  • It starts with a 7-day watch week: nothing changes, and AI Shield reports what it would have done.
  • Then choose Remove & review or Remove & strike, which uses your strike ladder. Harassment always goes to a moderator.
  • Every action explains itself, and Wrong call undoes it and posts the message back.

An example AI Shield entry in Remove & review mode.

05Strike ladder

Repeat offenders climb a ladder you set.

  • The default ladder
Repeat offenders climb a ladder you set.
Strikes in 7 daysWhat happens
1 strikeA warning.
2 strikesA 1-hour timeout.
3 strikesA 24-hour timeout.
4 strikesA kick.
5 strikesA ban.

Change the steps and the window (1 to 365 days) on the dashboard. A voided case stops counting.

Free vs Premium

Safety is free. On every plan.

See all plans and prices
AutoMod on Free and Premium
FeatureFreePremium
Up to 100 rules, the five ready protections and the link filterIncludedIncluded
Strike ladder, exemptions and member DMsIncludedIncluded
AI Shield150 AI credits a month1,500 AI credits a month

AutoMod questions

Quick answers.

What owners ask before they turn on AutoMod.

01Does Guardy Bot replace Discord AutoMod?

It runs alongside it. New rules run on Guardy Bot, and you read and edit them on the dashboard. Rules it handed to Discord's own AutoMod earlier are taken back by Guardy, and the dashboard says so until Discord confirms.

02What can an AutoMod rule detect?

Keywords, word lists, patterns, and spam such as message floods, duplicates, mass mentions, caps, emoji, reactions and spoilers. Links have their own link filter. Each rule can skip the roles and channels you choose.

03What happens when a rule matches?

Guardy Bot takes the action you set: delete, warn, time out, kick or ban. The member gets a private DM with the reason, the case lands in your mod log, and a rule that counts as a strike moves repeat offenders up your ladder.

04Does the link filter delete links straight away?

No. The link filter is off until you turn it on, and then it only records a case for moderators. You choose when it may delete messages or count strikes.

05Can AI catch scams that AutoMod rules miss?

Yes, if you turn on AI Shield. Your rules always run first; AI Shield then checks the messages they find suspicious but cannot decide on. It is included on every plan, starts with a 7-day watch-only week, and shows a reason and a Wrong call button for every action.

Protect the conversation

Write the rule once. Know exactly what it does.

Add Guardy Bot, turn on the five ready protections, and shape the rest around the way your server talks.