Moderation

The Moderation helper gives your staff the usual tools, keeps a case log you can review in the dashboard, and lets you decide which roles can use which action.

Turn it on

1. 1Open Moderation in the rail and flip it on.
2. 2Set a mod-log channel so every action is recorded.
3. 3(Optional) Set a mute role for /mute, or rely on Discord's native timeout.

Common commands

/moderation warn   user:@Member reason:Spamming
/moderation mute   user:@Member duration:30 unit:minutes
/moderation kick   user:@Member reason:Repeated rule breaks
/moderation ban    user:@Member reason:Raid
/purge contains    text:discord.gg count:50

Who can use what

In the module's Action permissions section you can allow specific roles per action (ban, kick, timeout, warn, nickname, purge). Leave a row empty to fall back to Discord's permissions; admins always bypass.

Recommended baseline

• Warn - allow junior moderators or helpers.
• Timeout / mute - allow moderators.
• Kick / ban / purge - restrict to senior moderators or admins.
• Nickname changes - allow only staff who handle safety or support.

Raid protection

A raid is many accounts arriving at once. That is the opposite problem from the anti-nuke guard below, which is about one account that already has power.

QREATE does not lock your server just because people are joining fast. A server that gets linked somewhere popular looks exactly like an attack to a simple counter, and locking it down at the moment it finally grows would be the worst thing we could do. So the join rate only opens the question. The answer comes from what the arriving accounts look like: how new they are, whether they have an avatar or a display name, whether their usernames share a stem (the user1234, user5678 pattern), and whether verification already flagged them. You set the weight of each signal, and raid mode trips only when the arriving burst averages enough suspicion.

When it trips, everything happens in one motion: new arrivals get a quarantine role, the server verification level goes up, the channels you chose get locked, and your team is pinged. Everyone already in the burst is caught too, because the raid started before we noticed it.

Then the part that matters most. The recovery screen lists everyone caught with their score and which signals they matched, and you sort them out in bulk: approve, kick or ban. Approving simply takes the quarantine role back off. One button reopens the server and restores exactly the verification level you had before.

Escalation engine

Discord's filter judges one message and stops there. The Escalation engine keeps a score per member instead: each thing it catches adds points, the score bleeds away while they behave, and crossing a level does something about it. Someone who shouts once is not a spammer. Six identical messages in eight seconds is.

Eleven filters can add points, each with its own threshold and weight: message speed, the same message repeated across channels, mention floods, invite links, links outside your allow list, capitals, emoji floods, unreadable text, attachment floods, walls of lines, and one person posting across many channels in seconds. The ladder maps score to action: delete, warn, timeout, kick, ban.

• Every automatic sanction writes a normal case with a case number, so it sits in the same log as your staff's actions and can be undone the same way.
• Points decay at a rate you set, so a talkative regular cools off instead of accumulating their way into a ban.
• Exemptions cover roles, channels, the server owner, and members you trust by how long they have been here or how much they have posted.

Word and spam filters

Under Word and spam filters you build Discord's own AutoMod rules without leaving QREATE. Discord checks every message itself, which is faster than any bot can be, and QREATE writes each block into your logging channel so you can see what the filter is actually doing.

• Words - a list of blocked words, one per line. A * matches any run of characters, so discord.gg/* catches every invite. There is an allow list for words your community uses innocently, and regex patterns if you want them.
• Discord's lists - the profanity, sexual content and slur lists Discord maintains, so you do not have to write them.
• Spam - Discord's own spam detection. Nothing to configure.
• Mentions - block a message that mentions more than N different people, with optional detection of coordinated mention raids.

For each rule you pick what happens: block the message (with your own notice to the member), post an alert in a staff channel, or time the member out. Roles and channels can be exempted per rule. Four starter packs (scam bait, invite links, slurs and profanity, mention flood) each open a normal rule you can read and change before saving.

Anti-nuke guard

This is protection against one member with power tearing the server down, usually a staff account someone else got into. It is not raid protection: a flood of new accounts is Verification's job.

QREATE watches the audit log and counts what a single member does in a short window: deleting channels or roles, creating them in bulk, granting dangerous permissions, banning or kicking in bulk, pruning members, creating webhooks, deleting emoji, changing server settings. For each of those you set how many times, within how many seconds, and what should happen: alert the team, take that member's roles away, or take the roles and lock the server.

The alert names who it was and carries two buttons. Give the roles back restores exactly what was removed. It was me, stand down leaves that member alone for the next 10 minutes, for when the burst was legitimate housekeeping. Both need Manage Server, and the same list with the same restore button sits on the Moderation page.

Trusted roles, trusted member IDs and the server owner are skipped. QREATE never reacts to its own actions, so bulk work you run from the dashboard is fine.

• It is off until you switch it on, per server.
• Discord sends audit records with a short delay and in batches, so it reacts within seconds rather than instantly. That is also why the default windows are generous.
• It cannot undo a deletion. Channels deleted before it stepped in are recovered from the 30-minute restore list; messages are never recoverable.

Lockdown commands

Temporarily freeze a specific channel or the entire server during a raid or emergency.

/lockdown server:false duration:30 unit:minutes reason:Spamming
/lockdown server:true duration:2 unit:hours reason:Mass spam in channels
/unlockdown server:false
/unlockdown server:true

Private staff notes

Add, view, and manage private staff-only notes on members. These notes are completely hidden from the member.

/notes setnote user:@Member note:Suspicious behavior in general
/notes view user:@Member
/notes removenote note_db_id:db-note-uuid
/notes clearnotes user:@Member

Purge examples

/purge contains text:discord.gg count:50
/purge user user:@Spammer count:25
/purge bot count:100

If an action fails

1. 1Check QREATE's role position. Discord blocks actions against roles above the bot.
2. 2Check the exact Discord permission: Moderate Members, Kick Members, Ban Members, Manage Messages, or Manage Nicknames.
3. 3Check whether the target member is the server owner; bots cannot moderate owners.
4. 4Check Logging or the case log for the failure context.