Server Assistant

Most Discord mod bots just react to messages.
This one understands your whole server.

Daily health insight · AutoMod that learns your team · fair ban appeals · a tamper-proof audit trail.

Free to start — the full moderation toolkit is free forever.

⚡ What makes Server Assistant different

Most bots are rule-runners: a message breaks a filter, the bot reacts. Server Assistant adds an intelligence layer on top — here's how.

🩺 Whole-server health insight

Every other bot waits for a rule to break. Pulse posts a health report to your staff channel — daily, weekly or monthly, your call — and flags the slow-burn patterns humans miss.

How it works

Messages, joins, AutoMod activity and warnings, each with a week-over-week trend — like a trickle of joins sitting just under your raid threshold. Set the cadence (or turn it off) in /settings → 🔔 Notifications. Read-only: it never acts on its own, it just gives your team x-ray vision.

Server AssistantApp9:00 AM

🩺 Daily Pulse — last 24 hours

Your server looks healthy. One thing worth a glance. 👇

Messages

4,812 · ▲ 6% vs avg

New members

37 joins · ⚠️ above baseline, below raid threshold

Read-only insight · never auto-actions

📩 Fair ban appeals

A ban shouldn't be a silent door-slam. The member is DM'd exactly why, and one reply opens a staff appeal ticket — so you decide with context, not vibes.

How it works

The appeal ticket has one-tap Unban, Deny, and Research — which runs an AI report on the member's last message. No DM noise, no per-incident tickets, no AI token spent unless you press Research. (A ThreatNet auto-ban is the exception — it's silent and opens no appeal ticket; that automated decision is contested via the web portal.)

Server AssistantApp#staff-log

📨 Ban appeal

"I was out of line and I'm sorry — I've read the rules properly now and won't do it again."

Ban reason

Repeated harassment after two warnings

✅ Unban🚫 Deny🔎 Research

🤔 AutoMod that asks before it over-reacts

Keyword filters punish the innocent — the classic "I love visiting Scunthorpe" problem. Borderline hits get an AI confidence check first.

How it works

Clear violations act instantly; anything doubtful is left up and handed to staff with one-tap Delete & warn or Allow. Less collateral damage, less mod drama — and only borderline messages cost tokens.

Server AssistantApp#staff-log

🤔 AutoMod — needs review

Flagged by word_filter · 32/100 confidence — below your threshold, left up for you.

AI note

Likely false positive — UK place name, not a slur

honestly I love visiting Scunthorpe in the summer

🗑️ Delete & warn✅ Allow

🧠 AutoMod that learns your team

Off-the-shelf filters are one-size-fits-all. Self-trained AutoMod studies the calls your staff actually make and proposes rules tailored to your community.

How it works

It clusters what your team deletes and reports, then suggests a server-specific rule for your approval. Nothing is ever auto-enforced — every rule is a suggestion until you accept it.

Server AssistantApp#staff-log

🧠 Suggested AutoMod rule

Your team removed 11 messages matching this pattern in two weeks. Want me to start catching it?

Proposed filter

crypto pump · "guaranteed 10x" · DM-for-signals spam

Learned from your staff · never auto-enforced

✅ Add ruleSkip

🛡️ A tamper-proof audit trail

Every action is recorded — even bans done by hand in Discord — in an encrypted log staff can't edit or erase, and the owner is pinged if anyone tries to delete an entry.

How it works

Native right-click bans/kicks/timeouts are captured like bot commands; the record is encrypted with the same key protecting your credentials; deleting a log-channel entry alerts the owner while the encrypted copy survives. No other moderation bot ties native-capture, encryption, and tamper alerts together.

Server AssistantAppto the owner

🛡️ Tamper alert — log entry deleted

Someone removed an entry from #server-assistant-log. The encrypted record is intact — here's what they tried to hide.

Deleted by

@rogue-mod

Hidden action

ban @user — "spam" (done by hand in Discord)

Owner notified · original preserved, encrypted

🕵️ Catches ban-evading alts

Repeat-offender detection fingerprints the people you ban and auto-stops their alt accounts at the door — and Threat Score turns raw behaviour into an explainable risk level.

How it works

A fresh account reusing a banned user's avatar or near-identical name is auto-banned and reported (weaker matches flagged). Threat Score is time-decayed, shown to staff in profiles — visibility first, with extensive safeguards before any action.

Server AssistantApp👤 Info → Threat Score

🛡️ Threat Score — @driftwood

🟠 Elevated · 58/100 — an explainable read of infraction history, weighted by severity and decaying over time.

Recent history

3 warnings · 1 timeout · 0 bans

Advisory view · free for every server

🌐 Knows the raiders before they arrive — ThreatNet

ThreatNet, our Cross-Server Threat Network, shares severity-only abuse signals across every protected server — so a scammer or raider with a serious, corroborated history elsewhere lights up on yours before they strike. A core, on-by-default safety net.

How it works

Servers contribute minimized, pseudonymous signals — counts, recency and a generic severity level — never the offence type, your staff's reasons, message content, or which server acted. The advisory band is free for every server; Premium adds the richer cross-server view. Individuals can opt out of profiling in the web portal, subject to a published safety exception. Premium auto-protect can ban a qualifying account on join; that ban is silent and opens no appeal ticket — it's disclosed in our Privacy Policy and stays contestable via the web portal.

Server AssistantApp👤 Info → Threat Network

🛡️ Threat Network — @driftwood

🟠 Elevated · corroborated across 3 servers — advisory cross-server safety signal.

Most recent

serious · 6 days ago

Severity-only · never which server acted

📸 What your staff actually sees

The same intelligence, in the moments it matters. Each panel is a to-scale mock of a real message — the situation on one side, what the bot does on the other.

A regular suddenly turns hostile in #general

Third flare-up with the same member today. Instead of guessing, a moderator right-clicks the message → Message Report.

What the feature does

Message Report reads ~20 surrounding messages and returns a 3–5 bullet AI summary — tone, likely rule broken, escalation pattern, suggested action — with one-tap Warn / Timeout / Kick / Ban. The bot never auto-acts; staff always decide, and the action is logged with the report attached.

Server AssistantAppright-click → Message Report

📋 Message Report

• Tone: hostile, escalating
• Likely breaks: rule 3 — personal attacks
• Pattern: third flare-up with the same member today
• Suggested: short timeout + a staff check-in

AI summary · reads ~20 messages of context

⚠️ Warn🔇 Timeout👢 Kick🔨 Ban

A day-old account drops a "free Nitro" link

Classic scam, posted while your team is asleep. AutoMod handles it before anyone sees it — and tells the user why.

What the feature does

Filter packs catch scams / slurs (IP-grabber domains preloaded); the message is removed and the user DM'd which rule fired plus their warning count. Staff are completely exempt, and repeat hits auto-escalate up your punishment ladder.

Server AssistantApp#staff-log

🤖 AutoMod · scam_link

User

@newbie_4821 · account 19h old

Action

Message deleted · auto-warning #1 · user DM'd

join for free nitro 🎁 discord-gift[.]ru/x

Logged · staff exempt · IP-grabbers preloaded

A mod tries to purge 200 messages

Powerful, irreversible — so it doesn't just run. The action routes up the chain for sign-off, and everything is logged.

What the feature does

Tiered approval: dangerous actions (like /purge) post an approval card to a higher tier — only the Owner can approve an Admin's request. Once approved it executes and is written to your audit trail with the approver recorded. No rogue purges, no silent mass-deletes.

Server AssistantApp#staff-log

⚖️ Approval needed — /purge 200

@admin-jess wants to purge 200 messages in #general. This routes up a tier for sign-off.

Requested by

@admin-jess (Admin)

Approver

Owner only

✅ Approve❌ Deny

🧰 The complete toolkit

Everything below is in the box. The free tier is a complete moderation suite; Premium adds your server's own power tools.

🖱️ Three ways to use it

Server Assistant · Features · Compare · Pricing · Roadmap · Support