Keep big communities healthy without babysitting them.

Start with roles, not rules

Almost everything else in moderation flows downhill from your role setup. A role is a bundle of permissions, and permissions on Vronify.chat are granular: posting, managing messages, managing members, and so on are separate grants, so you can hand a helper the ability to tidy a channel without also handing them the ability to ban. Per-channel overrides handle the cases a server-wide setting cannot express.

The honest edge case: overrides interact. A member with three roles and two channel overrides can end up with permissions you did not picture when you set each piece up individually. When it matters, check what that member can actually do in that channel before you rely on your mental model. And resist the urge to create a role per person; roles describe jobs, not people.

Slow mode is a pressure valve, not a punishment

Slow mode puts a pause between messages, per member, in the channel where you enable it. That single mechanic changes the physics of a heated room: nobody can flood, everybody gets read, and the loudest voice loses its main advantage. Use it during launches, big announcements, and any moment where the message rate outruns anyone's ability to read.

What slow mode does not do is settle an argument. It calms the flood, but if two members are genuinely in conflict, a moderator still needs to step in. And turn it back off when the surge passes: a permanent slow mode in a normal channel reads as distrust, and your regulars will feel it.

Automated filters are the night shift

Filters inspect traffic around the clock and stop the obvious junk, spam and abuse, before a human ever has to see it. That matters more than it sounds: the worst content in any community is the content your volunteers burn out on. A good filter setup means your moderators spend their energy on judgment calls, not garbage collection.

The trade-off is real and worth saying plainly: every filter balances missed abuse against false positives. A filter tuned to catch everything will eventually catch a regular making a joke, and nothing corrodes trust faster than automation punishing good-faith members. Start conservative, watch what the filters catch in the first weeks, and tune from evidence rather than fear. A filter should shrink your moderators' workload, never replace their judgment.

Member reports are your early-warning system

Reports let members flag a problem directly to your staff with the context attached, so your team is not reconstructing an incident from cropped screenshots and secondhand summaries. Faster context means faster, better decisions, and it means problems reach you while they are still small.

Two habits keep the system healthy. First, treat reports as signals, not verdicts: a report tells you where to look, and your staff decides what actually happened. Second, close the loop. A community that watches its reports vanish into a void stops filing them, and then you lose your early warning entirely. You do not have to agree with every report; you do have to make it visible that reports get read.

The audit log keeps everyone honest

Every moderation action lands in the audit log: who did it, to whom, and when. That record protects in both directions. Members are protected from a rogue moderator, because nothing staff does is invisible. Staff are protected from false accusations, because the record shows exactly what they did and did not do. When a dispute reaches you as the owner, you settle it from the log instead of from competing memories.

Bans that stay fair make banning easier

On most platforms a ban is a confiscation: the account, the messages, the DMs, all seized. Here, a ban only stops participation. The person you remove keeps their account, keeps every message they ever sent, and keeps the DMs they received, unless the sender removed them. If they want to leave entirely, they can delete their account at any time, which anonymizes it. That is the account ownership promise, and it applies to banned members too.

This is not just fairness for its own sake; it changes how your team behaves. When a ban is a proportionate tool instead of a nuclear strike on someone's digital life, moderators stop agonizing over it and use it when it is warranted. Firm decisions get easier when they are not also cruel ones.

Where moderation stops

Community channels are not private, on purpose: your moderators need to see public spaces to keep them safe, and the tools on this page exist for exactly those spaces. Direct messages are the opposite. DMs are private, which means we don't read them, and neither can you or your staff. The line is clean and it does not move: public spaces are moderated, private conversations are private. If that boundary matters to you, the privacy page covers the rest of it.

When you are ready to put all of this into practice, the moderation guide walks through a setup that keeps most problems small, and the owner handbook covers the decisions that sit above moderation entirely.

The engine is one system, built by us, running on the same infrastructure that delivers your messages, and each feature is its own toggle. It is not live yet; the bot engine page explains the whole design, and the release will be announced in the changelog. Everything else on this page works today.