Moderation & safety

Keep your server clean with a verification gate (Free), auto-moderation, an infractions log with warn escalation, an optional AI check, and alt-account detection.

HoneyGate hands you a full safety toolkit. The verification gate (Free) is your front door. Anti-nuke (Pro) stops a rogue admin before they burn the place down. Auto-moderation (Pro) filters the bad messages. An infractions log (Pro) keeps a permanent record with optional warn escalation, and alt detection (Pro) flags the accounts that are probably one person wearing five hats.

Verification gate (Free)

New members click a Verify button, pick up your verified role, and unlock the rest of the server. Set your channel permissions so an unverified account can only see the verify channel until it passes.

• Account-age check (optional). Turn away accounts younger than a number of days you pick. Raiders and scammers lean almost entirely on brand-new accounts, so this alone stops a lot of trouble.
• Alt screen (optional, uses Alt Detection). If the account has already been confirmed as an alt in your server, you can hold it for staff review or deny it outright.
• Raid lockdown (automatic). When a burst of accounts piles in at once (more than your threshold in a short window, 10 by default), HoneyGate locks the gate and parks every new arrival for staff review. It lifts once you say so, or on its own after the flood dies down. You can also trigger or release a lockdown by hand from the admin page.
• Every attempt is logged, and holds or denials post to a staff channel you choose.

Setup: turn it on, pick the verified role and the verify channel, optionally set the account age and alt action, then click Post panel. Your bot's role has to sit above the verified role so it can hand it out.

Anti-nuke (Pro)

This one guards against your worst day: a compromised or rogue admin, or a malicious bot, suddenly mass-deleting channels and roles, mass-banning members, or spamming webhooks. HoneyGate watches your audit log. When a single actor crosses a threshold of destructive actions in a short window, it responds and pings your staff channel. You choose how hard it hits: alert only, strip the offender's dangerous roles, kick, or ban.

• Always exempt: the server owner, the bot itself, and anyone you whitelist (by role, or by user/bot ID for bots you trust).
• Two requirements: the bot needs the View Audit Log permission, and its role has to sit above anyone it might need to stop. Without the height, all it can do is shout.
• Once your whitelist is set, start with Strip dangerous roles.

Auto-moderation

Auto-moderation reads messages and acts on the ones that break your rules.

Quick start: pick a protection level

You don't have to think through every setting. At the top of the auto-moderation page, pick a one-click protection level and it fills in sensible defaults:

• Relaxed. Light touch. It stops invite spam and message flooding but won't punish enthusiasm like all-caps or a pile of pings.
• Balanced (recommended). The sensible default for most communities. It filters invites, mass-pings, all-caps shouting, and spam, then times out repeat offenders for 5 minutes.
• Strict. Locked down for large or raid-prone servers, with tighter limits and a 10-minute timeout.

Picking a level fills the settings and switches auto-moderation on, so review them and click Save. Fine-tune anything afterward and it simply shows as a "Custom" level. A preset never touches your banned-words list, exempt roles and channels, log channel, or the AI check.

Fine-tuning

What you can filter:

• Invite links to other Discord servers.
• Links in general.
• Mass mentions over a limit you set.
• Excessive caps over a percentage you set.
• Spam, meaning too many messages in a short window.
• Banned words from your own list (whole-word, case-insensitive).

What happens on a violation: the message is deleted and, if you want, the member is timed out for a duration you choose. Everything lands in a log channel you pick. You can also exempt trusted roles and specific channels.

The optional AI check

Switch on an AI second pass and it catches what a plain word-filter misses: context-based toxicity, scams, and shilling. Two things matter here:

• It only flags suspicious messages for your staff to review. It never deletes on its own, so it fails on the safe side.
• It runs after the keyword rules, it's rate-limited, and if it's ever unavailable your normal rules keep working.

Infractions & mod log (Pro)

A permanent record of every moderation action against a member. Your paper trail.

• Log a case from the admin page: a note (record only), a warning, a timeout, a kick, a ban, or an unban. Add a reason and an optional link to evidence. Timeouts, kicks, bans, and unbans are also carried out on Discord for you.
• It records what you do in Discord, too. When a moderator bans, kicks, or times someone out directly in Discord, HoneyGate notices and adds it to the log automatically, so the history stays complete rather than just the panel actions.
• Per-member history. Click any member to see everything on file: total cases, how many active warnings they carry, and a full timeline.
• Automatic escalation (optional). Set thresholds like "3 warnings, timeout", "5, kick", "7, ban". Once a member racks up that many active warnings, HoneyGate acts on its own, and the most serious threshold they've crossed wins. You can pardon a warning so it stops counting.
• Optional log channel posts a tidy summary for every case, and you can DM the member the action and the reason.

Setup: open Infractions, pick a log channel if you want one, decide whether to DM members, and optionally turn on escalation with your thresholds. Then Save. For the bot to record Discord-side actions and carry out timeouts, kicks, and bans, it needs the matching permissions (View Audit Log, Moderate Members, Kick/Ban Members) and its role has to sit above the members it acts on.

Alt detection

Alt detection flags accounts that are likely run by the same person, using signals like shared network fingerprints, timezone and locale overlap, and VPN or proxy use.

Setting it up: turn it on, optionally enable the VPN/proxy check, set a confidence threshold for flagging, and choose a staff channel for alerts. You review each flag and either confirm or dismiss it.

Use judgment. These signals are probabilistic. People on the same home network, a campus Wi-Fi, or a shared VPN can look like alts. Review every flag before you act, and go easy on any auto-kick option.

Why it pairs with giveaways

Once alt detection is on, you can switch on "exclude alts" in giveaways so one person can't enter with five accounts.

Tips

• Start auto-mod gently (invites, spam, and a short banned-words list), then tighten as you learn your crowd.
• Always set a log channel so you can see what's being actioned.
• Keep your staff roles exempt so your moderators don't get caught by their own filters.