Ban appeals
With ban appeals on, anyone you ban gets a DM explaining exactly why, and can put their case in a single reply. That reply opens an appeal ticket in your log channel with one-tap Unban, Deny and context buttons. No audit-log digging, no per-incident channels, no “did you get my appeal?” chasing.
This page walks the flow, the free vs Premium difference, and the one exception: a ThreatNet auto-ban, which is silent by design.
Automatic once it’s on. Ban appeals are controlled from
/settingsand are on by default. There’s nothing for a banned member to install, they just reply to the DM.
Tags on this page:
FREE available on every plan · PREMIUM needs a Premium plan · ADMIN requires staff with the ban capability
Step through an appeal
Use Next › and ‹ Back to walk the flow: the ban DM, the member’s reply, the staff ticket and its buttons, the optional Premium back-and-forth, and the outcome.
The member is told why. They get the exact reason and appeal by replying to the DM, no button to find. If no reason was recorded, the DM reads “No reason was provided.” If their DMs are closed, the bot can't reach them and the appeal can't start.
One reply, submitted. They reply once and the bot confirms it's gone to staff. A second reply just gets “Your appeal is in with the staff team, you'll get a message here if there's an update.” On free, this single reply is the appeal; on Premium the bot can ask a short follow-up first (step 6).
The ticket. It lands in your log channel with the member's words, their ID and the original ban reason. Unban lifts the ban and DMs them; Deny closes the appeal and tells them the ban stands. Info, Warnings and Notes pull up their history privately, for context. More info is Premium (step 5). Only staff with the ban capability can press any of them.
Research AI. Press Research and the bot writes a short, private assessment of the member's most recent cached message. It's opt-in: it runs (and costs AI) only the moment you press it. If nothing was cached, it says so rather than guessing.
More info PREMIUM. On Premium, More info opens this form. Type a question, the bot DMs it, and the ticket pauses (“this card is paused; a fresh one returns when they reply.”). Their follow-up doesn't count against the one-reply limit, and a fresh ticket returns the moment they answer.
Multi-round appeals PREMIUM. The bot can ask one or two short follow-up questions before passing the appeal on, turning a one-line “sorry” into something staff can act on. The AI only facilitates, it never decides the outcome. Staff get the full back-and-forth plus a short, neutral summary.
The outcome. The member hears back in their DMs: an approval comes with the unban and a nudge to read the rules; a denial says the ban stands.
Free vs Premium
Ban appeals are free on every plan: the reason DM, the single-reply appeal, the staff ticket and its Unban / Deny / Research / Info / Warnings / Notes buttons are all included. Premium adds the guided, conversational layer.
On by default. A banned member is DM'd the exact reason and invited to send one reply, which opens a staff appeal ticket in your log channel. Staff get one-tap Unban and Deny, the Research AI report (opt-in, costs nothing unless pressed), and Info, Warnings and Notes for context. Only staff with the ban capability can act on it.
Adds a short, AI-facilitated back-and-forth: instead of taking a single reply, the bot can ask one or two clarifying follow-up questions to gather the full picture, then hands staff the whole conversation with a neutral summary. The AI only facilitates, it never decides the outcome. Premium also unlocks the More info button, so staff can ask the member a question mid-appeal without spending the member's one reply.
“Why was I actioned?” Separately, on Premium, the DM for a warning or other action (not a ban) carries a Why was I actioned? button. One tap and the bot writes a short, plain-English explanation of what the action was for, heading off confused replies to your staff.
Nothing quietly rots
An appeal nobody gets to shouldn’t sit open forever, and it shouldn’t vanish without anyone noticing either. Two things make sure of that.
SAi DMs a staffer before an appeal closes itself. Once an appeal has run past its deadline, SAi sends one staff member who can actually action it a direct message saying plainly that the card is about to close and exactly when — as a Discord relative timestamp, so it reads correctly in any timezone and stays correct however long the DM sits unread. It also says how long the appeal has been waiting and links straight to it in the portal. This is deliberately not the ordinary “been waiting about an hour, no rush” nudge: an overdue card gets its own wording, because telling a moderator to take their time over something that lapses today is worse than saying nothing.
- It’s sent to the people who could action the appeal, plus the owner.
- It respects the same opt-out as every other staff nudge — turn “Call me by DM” off in the portal and both the ordinary escalation nudges and these last-chance warnings stop together. There’s no way to be opted out of one but not the other.
- The portal feed entry and web push still fire; the DM is an extra channel, not a replacement.
- If DMs are down, the warning is simply missed — a card still closes on schedule.
Then it closes on a clock. A ban appeal (and a Minecraft ban appeal) is kept for 30 days — deliberately the generous end, because someone is waiting on an outcome. After that it’s warned as above, closes following a 24-hour grace period, and is deleted 90 days after closing. The full table, including the shorter window on cards that carry a message extract, is in the audit log guide.
The exception: a ThreatNet auto-ban
One ban does not trigger any of the above. When ThreatNet auto-protect (a Premium safety net) bans an account on join for a serious, corroborated abuse history across other protected servers, that ban is silent:
- the member is not DM’d a reason, and
- no appeal ticket is opened in your log channel.
This is deliberate and disclosed in our Privacy Policy. A ThreatNet auto-ban is an automated, cross-server decision, so it is not appealed through the in-server flow on this page. Instead it remains contestable via the web portal, where the affected person can challenge the automated decision and manage their profiling preferences.
Contestable, not unappealable. A ThreatNet auto-ban can always be challenged, through the web portal rather than a ban-appeal DM. See ThreatNet for how the cross-server signal works and what it does (and doesn’t) share.
See also
- Member experience, the welcome DM, verification and every other flow a member sees
- ThreatNet, the cross-server threat network and its silent auto-ban
- Privacy Policy, what’s collected, kept and shared, and how an automated decision is contested