Pre-launch

Discord raid cover

The raid ends. The questions start.

Anti-raid tooling is judged on the twenty minutes of the attack. It should be judged on the week afterwards, when somebody has to explain which of those decisions were right.

What you are actually facing

A raid is a schedule, not a surprise.

Coordinated joins are planned somewhere else and executed on your server in a few minutes. The defence that matters is not cleverness during those minutes. It is that the rules were already decided, and that what the rules did is still legible when everybody has calmed down.

How do anti-raid bots respond to a coordinated attack?

By recognising that a lot of accounts are behaving alike at the same time, and tightening the door while it lasts. That is the general shape across the category. The differences between products are mostly about how quickly they notice and how much they are willing to do without asking anybody.

The second half of that sentence is the one to think about. Speed is worth having, and the fastest possible response is a system with permission to permanently remove large numbers of people on a pattern match. That is also the response that turns a bad afternoon into a community you have to rebuild, and it is the reason the limits on this site do not have a raid exception in them.

The other direction

Anti-nuke is about your own side.

A nuke is not strangers at the door. It is an account that already has permissions deleting channels, stripping roles or mass-removing members, usually because it was compromised rather than because its owner turned.

What does anti-nuke protection actually protect against?

Damage done with legitimate permissions. That is what makes it a different problem from a raid: nothing about the actions is unauthorised, so a filter looking for outsiders sees nothing wrong. What gives it away is the shape and the speed of what a trusted account is suddenly doing.

It is also the case where a record earns its keep twice over. Afterwards you need to establish what was destroyed, by which account and in what order, both to put the server back and to work out which credential failed. A history that only lists outcomes tells you the first part and none of the second.

It does not stay on one platform

Both doors, one rulebook.

The same operators try Discord and Telegram, usually within the same day. Two bots on two platforms with no shared record means the second one meets the attempt as a stranger, and your moderators are the integration layer.

Can one anti-raid policy cover Discord and Telegram together?

Yes. Both communities run from one workspace on one Community Blueprint, while each platform and each channel keeps its own permissions. Tighten the response because of what you saw in one place and you have tightened it in both, without anybody having to remember the second tool on a day when they are already busy.

What it costs

The bill does not spike with the traffic.

A raid is, in billing terms, a large amount of activity arriving quickly. On usage-priced tooling that is an invoice. It is worth checking what your current stack does on that day before you find out.

Is raid protection included, or a paid add-on?

Included, at every volume and on every plan. These are never metered:

  • Live moderation
  • Scam and impersonation detection
  • Raid and anti-nuke protection
  • The incident queue
  • Appeals
  • Audit history
  • Sentiment and analytics
  • Setup

Servers, groups and seats are unlimited on every plan including the free one, so covering a second community is not a second charge. What separates the plans is how many workspaces you run, which is a number you choose rather than a number an attacker chooses for you.

Where the automation stops

What the AI is not allowed to do.

The reason to trust a moderation product is not how clever it is. It is what it is prevented from doing without you.

  • AI cannot ban a member on its own.
  • AI cannot apply a permanent sanction on its own.
  • AI cannot publish a change to your policy.
  • Deterministic low-risk rules act automatically. Everything above that line waits for a person.
  • Every action is reversible, and the reversal is recorded too.
  • Provider identities are stored as one-way pseudonyms, not as raw platform accounts.

Straight answers

Asked during the quiet week, not the loud one.

What is a Discord anti-raid bot?

A Discord anti-raid bot watches for a coordinated influx of accounts joining or posting together and responds before a server is overwhelmed. SwarmUp covers raids as part of moderation rather than as a separate tool, so what happened during one stays in the same record as everything else.

What is the difference between a raid and a nuke?

A raid comes from outside: many accounts arriving at once to flood, spam or harass. A nuke comes from inside, when an account with permissions deletes channels, mass-bans members or strips roles. One is a volume problem at the door; the other is a trust problem above it.

Will the bot mass-ban accounts during a raid?

No. AI cannot ban a member on its own and cannot apply a permanent sanction on its own, and a raid does not suspend that. Deterministic low-risk rules act automatically, and anything that would permanently remove people waits for a person to decide it.

What can I see after a raid is over?

The record of what happened, action by action, with the evidence and the owner attached to each one. Reversals are recorded too, so a batch of decisions made quickly under pressure can be reviewed calmly afterwards and corrected without the correction disappearing.

Do raids move between Discord and Telegram?

Often, yes. The same accounts and the same operators tend to try both doors, and a raid that starts in a Telegram group is frequently already queued against the Discord. SwarmUp runs both from one workspace, so the second platform is not meeting the attempt for the first time.

Does a raid make my bill go up?

No. Raid and anti-nuke protection is never metered, community growth is not metered, and only AI generation you ask for uses credits. A bill that follows volume would charge you for being attacked, which is the one moment a community should not be thinking about cost.

Where we are today

SwarmUp is not open to the public yet. Moderation is the core of the product and it runs today. Engagement and rewards are still in testing. Plans and prices are final, and nothing is charged before the workspace opens.

One email, when it opens

We will not waste the one email.

We are not naming a launch date until we can hold it. What you get for an address is a single message the day the workspace opens, and the same plain account of where the product stands on that day.