Discord raid tooling has gotten faster and more accessible over the past few years — what used to require real technical skill to build is now packaged into paid "raid panels" anyone can rent for a few dollars. That means the volume and speed of an attack has gone up while the skill required to launch one has gone down. The good news: the defenses that stop raids haven't fundamentally changed, and they don't require a specific bot to understand or start applying today.
This is a standalone, practical guide — the concepts here work regardless of what tooling you're running, followed by a section on how Novrix specifically implements each one if you're looking for a ready-made solution.
What a Raid Actually Looks Like
A raid is a coordinated wave of accounts joining a server in a short window, typically to spam, post shock content, harass members, or simply overwhelm moderators so real conversation and reports get buried. It can look like dozens of near-identical new accounts joining within seconds of each other, followed immediately by a flood of messages, mass pings, or NSFW content dumped across multiple channels at once.
The important thing to internalize: the defining signal of a raid is never a single account. It's the pattern — the rate at which accounts are arriving, how new those accounts are, and how coordinated their behavior looks once they're in. Every defense below is built around detecting or disrupting that pattern.
Defense 1: Verification Systems
Verification is the gate between "joined the server" and "can actually participate." It's the first and most fundamental layer, because a raid that can't post, react, or even see most channels immediately upon joining is a raid that's already been mostly neutralized.
Discord's built-in verification levels
Under Server Settings → Safety Setup, Discord offers native verification levels ranging from "must have a verified email" up to "must have a verified phone number and be a member of the server for a set amount of time before posting." This costs zero setup time and is worth having at a reasonable level (email-verified, minimum) on every server, all the time — not just during an active raid.
Server-specific verification (unverified role + gate)
A more targeted approach: new members get an "unverified" role automatically on join, which only allows them to see a single welcome/rules channel. They complete a short action — a button click, a captcha, answering a simple question, or reading and reacting to the rules — before a bot grants them a "verified" role that unlocks the rest of the server. This is more effective against raid tooling specifically, because raid tools are usually built to overwhelm Discord's generic, server-wide defaults, not a workflow that's unique to your specific server.
Balancing friction and conversion
Every extra step in verification loses some percentage of real new members who get impatient or confused. The servers that get this right keep verification to one clear, fast action, explain in plain language why it exists, and make the unverified state feel temporary and welcoming rather than like a punishment or a wall.
Defense 2: Anti-Join-Spam (Join Velocity)
This is the core automated detection layer: tracking how many accounts join within a rolling time window, and automatically triggering a response once that rate crosses a threshold you define.
Why velocity, not volume
A server that's genuinely growing might get 50 new members in a day — that's not suspicious on its own. But 15 accounts joining within 10 seconds of each other is a completely different pattern, regardless of your server's overall size or growth rate. Velocity (joins per short window) is the signal that actually distinguishes organic growth from a raid, and it's what a join-spam detector should be watching, not raw daily join counts.
Setting a sane threshold
Too sensitive, and a genuine event (a giveaway, a partnership shoutout, a viral post linking your server) triggers false positives that lock out real new members. Too loose, and a real raid slips through before it's detected. A reasonable starting point for most communities is somewhere around 10–15 joins within a 10–15 second window, with a wider, more lenient threshold (20–30 joins per 30 seconds) if you regularly run promotions or giveaways that cause legitimate join surges.
Defense 3: Account-Age Checks
Raid tooling relies on being able to mass-create or mass-rent Discord accounts cheaply, and those accounts are, almost by definition, new. Gating participation — or in stricter setups, gating the ability to join at all — behind a minimum account age directly targets that dependency.
How much friction is reasonable
A threshold of a few days (not weeks) is usually enough to filter out the bulk of disposable raid accounts without meaningfully affecting real users, since genuine new Discord users are very rarely joining a community with an account that's only hours or minutes old. Pushing the threshold much higher starts trading raid protection for lost real members, so it's worth tuning based on what your server actually experiences rather than defaulting to the strictest possible setting.
Pairing account age with action, not just visibility
Account-age checks are most effective paired with an automatic action — auto-kicking or quarantining accounts below your threshold — rather than just flagging them for a moderator to review later. By the time a human reviews a flag, a fast-moving raid has often already done its damage.
Defense 4: Server Lockdowns
Lockdowns are the blunt, high-confidence response: when a raid is confirmed or strongly suspected, temporarily restrict what the entire server (or specific channels) can do, buying your team time to assess and respond without new damage accumulating in the meantime.
What a lockdown typically restricts
Common lockdown actions include disabling new invites, raising the server's verification level to its strictest setting, restricting @everyone's ability to post in public channels, and in severe cases, temporarily closing the ability to join the server entirely. None of these need to be permanent — a lockdown is meant to be a short, sharp pause, not a new steady state.
Manual vs. automatic lockdowns
A bot with join-velocity detection can trigger a lockdown-equivalent response automatically (kicking or banning burst-joined accounts, for example) the moment a raid is detected — faster than any manual intervention. But manual lockdown steps (raising verification level, disabling invites by hand) remain useful as an immediate backup any admin can execute in seconds, with or without a bot.
What to Do During an Active Raid
- Raise verification level to the strictest setting immediately (Server Settings → Safety Setup) — this alone stops most unverified raid accounts from posting further.
- Disable invite creation, or delete active invite links, so the raid can't pull in reinforcements mid-attack.
- If you have automated join-velocity detection, confirm it's actively kicking/banning burst-joined accounts; if not, or if you're doing this manually, start banning accounts that joined in the last few minutes as a batch.
- Post a short status update in your announcement channel so real members know what's happening and aren't left guessing — panic and confusion often do more damage to a community's trust than the raid content itself.
- Once join velocity has returned to normal, gradually relax the strictest settings rather than leaving maximum lockdown in place indefinitely.
After the Raid: Cleanup and Review
Once things are stable, review your join/leave and moderation-action logs to confirm exactly what happened — how many accounts joined, what they posted, and what actions were taken. This is also the point to check for false positives: real new members who got caught in an aggressive automatic response and need to be manually restored. Finally, write down (or update) a raid response plan while it's fresh: who raises verification, who handles the announcement, and what your actual thresholds should be going forward based on what you just observed.
How Novrix Implements These Defenses
If you'd rather have these defenses running automatically instead of executing each step by hand, this is exactly what Novrix's Anti-Raid module is built around. It tracks join velocity against a configurable threshold and time window (defaulting to 10 joins per 10 seconds), and automatically kicks, bans, or locks the server the moment that threshold is crossed — no one needs to be watching for it to trigger. Account-age gating is built in as a companion setting, auto-filtering accounts younger than a configurable threshold (7 days by default). You can set a dedicated alert channel so your mod team is pinged the instant Anti-Raid fires, and everything it does shows up in the same case history and action logs as your manual moderation, through the dashboard's Anti-Raid page.
For the verification-gate approach described above, Novrix's onboarding and role tooling let you configure an unverified-role flow specific to your server, and its AutoMod module catches the message-spam wave that often follows a raid attempt — the two systems are designed to work together rather than as separate, disconnected tools. See the full documentation for setup details, or the complete Discord server security guide for how anti-raid fits alongside AutoMod, verification, and logging as one system.
Quick-Reference Checklist
- Keep at least email-verification enabled on Safety Setup at all times, not just during raids
- Set a join-velocity threshold appropriate to your server's normal growth pattern
- Gate new accounts with a minimum account-age of a few days
- Know your manual lockdown steps (raise verification, disable invites) before you need them
- Set an alert channel so your team is notified the instant automated raid detection triggers
- Review join/leave and moderation logs after any incident to catch false positives
- Write down your raid response plan in advance — deciding roles and steps mid-raid wastes critical time
Frequently Asked Questions
What is a Discord raid, exactly?
A raid is a coordinated wave of accounts joining a server in a short window of time, usually to spam, post shock content, harass members, or overwhelm moderators. The defining signal isn't any single account — it's the pattern of how fast accounts are arriving together.
What's the single fastest thing to do during an active raid?
Raise your server's verification level immediately under Server Settings → Safety Setup, and lock down invite creation. Both are manual, bot-independent actions you can take in seconds while an automated system (if you have one) handles the account-level response.
Does account-age gating stop real new users from joining?
It adds a small amount of friction, but a reasonable threshold (a few days old, not weeks) filters out the disposable accounts raid tools mass-create without meaningfully blocking genuine new Discord users, who are very rarely using an account that's only minutes old.
Can I fully automate raid response, or do I still need a human?
You can automate the fast, pattern-based response (kicking or banning accounts that joined in a burst, locking the server), and that automation is what buys you the time. But a human should still review what happened afterward — automated systems occasionally have false positives during legitimate growth spikes, like a giveaway or partnership push.
Conclusion
None of these four defenses — verification, join-velocity detection, account-age checks, and lockdowns — work perfectly in isolation, but layered together they cover the raid lifecycle from before an account can even post, through detecting the pattern in real time, to containing the damage if something does get through. The specific tooling you use matters less than having all four layers in place and a plan for who does what when they trigger. Want these running automatically? See how Novrix's Anti-Raid works, or add it to your server for free.