Protector documentation

Quick start

1. Add the bot to your server with the Add to Discord button.
2. Move the "Protector" role to the very top of the role list (Server Settings, Roles). A role can never act on a role above it: without this, anti-nuke, restore and sanctions fail.
3. Run /autoconfig, pick your language and a protection level (1 light, 2 standard, 3 strict). Protector enables and tunes every module, creates a log channel and a verification gate. You can run it again: anything that already exists is reused, never duplicated.
4. Check with /diagnostic that every protection has the permissions it needs.

Configuration

The /config panel

/config opens an interactive panel in a message only you can see. Each module has a button to turn it on or off, forms for thresholds and durations, and menus to pick channels and roles. Everything is saved instantly, per server.

Only members with the Manage Server permission can open the configuration.

Server language

The Language button sets the language Protector writes in on the server: English, French, Spanish or German. The /informations command switches the language for you alone, while you read.

Protected modules and security role

An administrator planning a betrayal would only need to turn the protections off before attacking. The most important systems are therefore reserved to the server owner:

• anti-raid, anti-nuke, anti-spam, anti-webhook, bot quarantine and coordinated raid;
• the list of staff roles, which are exempt from anti-spam.

Other managers can see these settings but cannot change them. In /config, General, the owner can choose a security role allowed to change them. Only the owner can give this role: if an administrator or a bot grants it to someone, Protector removes it at once and warns the owner in a private message.

Web dashboard

On protector-bot.com, log in with Discord to see the servers you own or where you have Manage Server, and turn each module on or off in one click.

The site never touches your server. It stores a request, which Protector picks up within seconds. The bot then checks on Discord that you are still a member, still have the permission, and that the module is not protected, then applies or refuses. Every change is recorded in your log channel with the mention "via the web dashboard".

Everything can be set from the site

On a server page, a module's Configure link opens its settings: thresholds, durations, percentages, texts, channels and roles. It is exactly what /config offers on Discord, with the same limits and the same protections. A bar appears at the bottom of the screen as soon as a value changes: nothing is sent until you click Save.

The site checks what it can right away (ranges, format, channel type), but Protector always decides: when applying, it checks every value and every permission again on Discord. A request goes through whole or not at all: if one field is refused, its reason appears right below it and nothing is written.

The general settings (language, log channel, staff roles, security role) and the multi-server network have their own page, at the top of the module list. Protected modules stay reserved for the owner and the security role: other managers see them read-only.

Welcome card background image

This is the only setting that lives only on the site, in the Welcome module. Choose a PNG or JPEG image (8 MB maximum, at least 440 × 130 pixels): Protector downloads it and scans it with its anti-NSFW filter before using it, more strictly than in channels, suggestive nudity included. If the scanning model is unavailable, the image is refused: no image is ever put in place without being scanned.

An accepted image is center-cropped to the card format (880 × 260) and re-encoded without its metadata, then stored by the bot. A refused image is deleted from the site at once; nobody but the person who sent it sees it before it is accepted. The Back to the default background button removes it from both the bot and the site.

Protection systems

Anti-raid

Detects an abnormal wave of joins over a short period and puts the server in lockdown: the verification level goes up, and accounts that are too new are kicked or banned for the duration of the raid. Everything is adjustable: join threshold and window, minimum account age, action, duration.

The lockdown is saved: restarting the bot does not interrupt it. At the end, the original verification level is restored, unless the staff changed it in the meantime, in which case their choice is kept.

Anti-spam and giant messages

Blocks flooding, repeated identical messages, mass mentions and invite links. Apps added to a profile (user apps) are watched too: when they spam, the author is kicked, the app banned when possible and the messages deleted. The same message spread across several channels also alerts the Protector team.

Protector also spots giant messages that flood a channel: "#" headings, very wide characters such as "﷽", empty or invisible lines. It estimates the on-screen height of the message and treats it as spam above 50 lines, or from twenty empty lines. The height is set in /config, Anti-spam.

If the author of a mention spam cannot be muted (missing permission, role too low), the owner is warned privately.

Coordinated raid

Anti-spam reasons per member. Twenty accounts each posting two identical messages therefore stay under all of its thresholds. This module watches a time window instead: if several different members post very similar content within a minute, the messages are deleted and the participants are muted while the staff takes a look. The comparison is done locally, and the server's staff is always exempt.

Anti-nuke and restore

If a member or a bot chains destructive actions (deleting channels or roles, mass bans, webhooks), Protector neutralizes them, alerts the owner, then restores the structure from a recent backup: channels, roles, permissions, positions, name and icon. Your configuration is automatically linked to the rebuilt channels.

Attacks that delete nothing count too:

• Prune: counts as many kicks as members removed.
• @everyone receiving Administrator, Ban, Manage Roles…: Protector reacts the first time and removes those powers.
• Private channels opened to everyone: each one counts, and Protector closes them again.
• Vanity invite changed or removed: the owner is warned at once, to claim it back before another server takes it.

Sanctions are graduated for humans: the global blacklist only applies beyond a certain number of destructive actions. Below that, it is a local sanction and a restore. A raider bot is always blacklisted.

Private channels and November 16, 2026

On that date, Discord stops giving bots the name and permissions of channels they cannot see. Protector records the state of every channel as it goes: after a nuke, a private channel is rebuilt with the name and permissions it knew. A channel whose permissions it never knew is not rebuilt incorrectly: the summary says so. Before the deadline, the owner receives the list of affected channels privately.

Anti-phishing, anti-webhook and anti-bot

• Anti-phishing: deletes scam links (fake Nitro, account stealers, lookalike domains) and sanctions their author.
• Anti-webhook: when a webhook posts spam or phishing, the webhook itself is deleted.
• Anti-bot: bots flagged as malicious are banned as soon as they join. Trusted bots can be allow-listed.

New bot quarantine

Enabled by default. Every added bot loses its permissions and extra roles until the owner decides, in a private message, to approve or kick it. While it waits, any attempt to give it roles or permissions back is undone.

On approval, Protector gives its permissions back. Discord only lets it give the ones it has itself: those it cannot give back are named, along with the role to add them to. Bots approved by the Protector team are not affected.

New bots. A bot whose account is less than 7 days old arrives with a warning in the message sent to the owner, and an extra Ban button. A bot younger than 30 days then stays under watch: on its first destructive action (deleting a channel, role, webhook or emoji, banning, kicking, pruning, adding another bot, giving powers to @everyone, opening a private channel to everyone), its permissions are removed, what it did is restored where possible, and the owner decides in DMs: allow, kick or ban. Protector never bans it on its own. A bot on the server whitelist is not affected.

NSFW filter

Analyzes text, with a list that resists workarounds ("p.o.r.n", leetspeak) without blocking innocent words, and images, with a model that runs locally. GIFs shared from Tenor, Giphy or Imgur are checked across several frames. You can allow specific channels and choose the action: delete, or delete and time out.

Honeypot channel

A channel where nobody has any reason to write, and which says so plainly. Hacked accounts and raid bots post everywhere without reading: a message in this channel gives them away. Protector acts immediately (softban by default, or ban, kick, timeout) and deletes their recent messages, up to 7 days back. After a kick or a softban, a link to come back can be sent to them: useful for a hacked account once its owner has recovered it. Staff, the owner, the whitelist, bots and webhooks never trigger the trap. Turn it on in /config, Honeypot channel; Protector can create the channel itself.

Risk score

For every join, Protector computes a score from 0 to 100 combining account age, network reputation (already blacklisted elsewhere?), default avatar and how closely the name or avatar matches known raiders. The result is logged; an automatic action above a high score is possible, disabled by default.

Staff guard

When a role with sensitive permissions (Administrator, Manage Server, Ban, Webhooks…) is given to an account considered risky (too new, recently joined), Protector alerts the owner and removes the role when possible.

Compromised staff account. Filters spare the staff, which left a stolen account free to act. Protector therefore watches signals that have no reason to come from a legitimate moderator: phishing link, burst of mentions, flood of messages. Beyond a threshold, their roles are removed and remembered, they are muted, and an alert goes to the owner. Access can be restored at any time.

Anti-impersonation

Detects members who copy the name or avatar of the owner or the staff, when they join and after a name change. The action is chosen in /config: alert or kick.

Native Discord features

Protector also drives the security features built into Discord itself. They act at the source: before a message is posted, before an invite is used, before a DM is sent.

Invites and DMs paused during a raid

When the anti-raid locks the server down, Protector asks Discord to pause invites and direct messages between members until the lockdown ends (24 hours at most, a Discord limit). Members see Discord's official "unusual activity" banner. This is the only answer to raiders who send scam DMs to your members: no bot can see or block a private message.

The pause is lifted when the lockdown ends. A pause set by your staff by hand is never lifted by Protector. Setting: Pause invites and DMs in Anti-raid (on by default). Protector needs the Manage Server permission.

Discord AutoMod shield

Protector reads a message and then deletes it: it always acts after the message is posted. With the Discord AutoMod shield (Anti-spam module), Protector also keeps native AutoMod rules up to date from its own configuration:

• Banned words and invites: the words of the word filter and Discord invite links;
• Mass mentions: the mention limit of the anti-spam, with Discord's mention raid protection;
• Spam: content Discord itself recognizes as spam.

Discord blocks these messages before they are posted, with no delay, and keeps protecting even if Protector is offline. Staff roles and whitelisted roles are exempt, as are the channels excluded from the anti-spam and the word filter. An alert goes to your log channel.

The rules are named "Protector · …" and Protector only ever edits its own: your AutoMod rules are never touched. Discord allows one spam rule and one mention rule per server: if your staff already has one, Protector leaves it in place. Turned off, the shield removes its rules. /autoconfig turns it on.

Discord onboarding

/autoconfig and the server builder set up Discord's native onboarding: the default channels shown to newcomers and, with the builder, a "Pick your roles" question with the server's cosmetic roles.

• Discord only offers onboarding on Community servers (Server Settings, Enable Community), with at least 7 channels visible to everyone, 5 of them open for writing. Otherwise Protector tells you what is missing.
• /autoconfig never replaces an onboarding your staff already set up. The server builder, which rebuilds the whole server, replaces it.

Linked roles: "Verified by Protector"

Discord lets an app provide requirements for linked roles. Protector offers two:

• Verified by Protector: the account is linked to Protector and is not on its global blacklist;
• Account age (days): the Discord account was created at least this many days ago.

For a server: Server Settings, Roles, pick a role, Links, Add requirement, Protector. Members who meet the requirements claim the role themselves from the server menu. It works even on servers where Protector is not installed.

For a member: the link is made once, on protector-bot.com/linked-roles, and follows you to every server. If your status changes (blacklisted, for example), Discord is updated automatically. You can unlink at any time.

Member verification

New members are quarantined: they only see the verification channel. Three modes are available in /config, Verification:

• Button: one click on "Verify" grants access.
• Image captcha: a code is posted with a mention of the member, who types it in the channel. Right code: the messages are deleted and access is granted.
• Website (Premium): the button gives a private link to protector-bot.com. The member signs in with Discord and Protector spots alternate accounts by device. An account linked to a member banned from the server is kicked (or banned, or left to the staff, your choice); another probable alternate account stays in quarantine and the staff decides with the alert's buttons. With Premium Plus, a connection through a VPN, a proxy or Tor is also spotted. Settings: dashboard, Verification, Web verification.

Setup: /verification-setup, with the quarantine role and the verification channel. You can also set a role added automatically after verification.

If the verification message is deleted, it comes back on its own when the next member joins. Roles given by other bots to a member not yet verified are removed, then given back once they verify. After setup, Protector checks what the quarantine role can actually see and names any channel still visible.

Moderation and records

• /warn /mute /kick /banSanctions, each requiring the matching Discord permission.
• /tempbanTemporary ban, lifted automatically.
• /unmute /unbanLift a sanction.
• /historyA member's record on this server.
• /clearBulk delete up to 500 messages: all, a member's, or those containing a word.
• /slowmodeChannel slowmode.
• /lockdown /unlockEmergency lock of the whole server.
• /panicExtended emergency: lockdown, revoke every invite and filter joins.

Every sanction is saved in a per-server record. Sanctions made outside the bot, through the Discord audit log, are captured too with their reason. With /clear, pinned messages are kept; beyond 14 days, Discord forbids bulk deletion and the reply says so.

Word filter

Blocks a list of words you define, ignoring case and accents. Enable it in /config and manage the list with /wordfilter.

Moderation improvement

When a filter acts on a message, Protector may submit that content to its team to check that the action was justified. Authors are replaced by an irreversible pseudonym and the content is erased after 90 days. It can be turned off at any time in /config, Improvement. Details are in the privacy policy.

Reports

• Report message (right-click a message, Apps): a form sends the message to your server's report channel, where the staff can sanction or dismiss.
• Report user (right-click a member, Apps): a report with reason, links and images goes to the Protector team, who reviews it without joining your server.
• Report to Protector (right-click a message): sends a scam to the team, even outside protected servers.
• On the support server: open a report on the support server, with a detailed description and your evidence (screenshots, videos, logs).

An account reported several times and confirmed by the team can be added to the network blacklist, which protects every server at once.

Event log

Disabled by default. Once enabled, you choose which categories to log and a default channel; each category can have its own channel. Logs are sent through a webhook specific to each channel.

Reading a change log

Every change is written before, then after. The "By" field tells three situations apart: a named member (it was them), "Missing permission" (give Protector View Audit Log), or "Not attributable" (Discord wrote no audit entry).

Who deleted a message

• Protector, with the reason ("Protector, anti-spam"): the bot records its own deletions.
• A named member: the deletion comes from the Discord audit log.
• "No trace, most likely the author": Discord wrote nothing. It is very probably the author, without certainty, and the log does not claim otherwise.

Backup and restore

Reserved to the owner, and independent from the anti-nuke's automatic backups.

• /backup createCaptures the full structure: roles, channels, permissions and recent text messages.
• /backup listLists your backups.
• /backup deleteDeletes a backup.
• /restoreShows a backup's content, lets you choose whether to restore messages, then restores.

A backup is created automatically when the bot joins. Two backups at most per server, five for a Premium server, ten with Premium Plus.

Network blacklist and appeals

When a user reaches a very high offense level, or during a confirmed nuke, they are globally blacklisted: banned from every server Protector protects, and blocked if they try to join one. They are warned by private message beforehand. The blacklist power is reserved to the Protector team: server administrators do not have it.

You can still choose to let them in. If they come back to your server, Protector removes all their access and sends you a message with two buttons. Without an answer within five minutes, they are banned. Either way, the account stays barred from the bot's commands.

The only place where a blacklisted account keeps a voice is the support server, where /appeal asks for a review.

Premium and Premium Plus

All the protection described above is free, including AI moderation of flagged messages. The plans add optional analysis and tools; the full comparison and subscription links are on the Premium page.

Protector in your colours (Premium): on your server, the bot takes the nickname, avatar, banner and bio you choose in Branding, without changing anywhere else. Images go through the anti-NSFW filter, a nickname imitating Discord or another bot is refused, the bio keeps a "Powered by Protector" line, and everything returns to the Protector look when the plan ends.

Community tools: the level card background, 10 voice creator channels, notifications for all your accounts and 25 auto responses with Premium; unlimited creator channels and 50 auto responses with Premium Plus. Premium also adds advanced moderation statistics; Premium Plus adds sanction appeals and an application form of your own.

In /config, a module reserved to a plan is still shown with a star and a padlock, but cannot be set without the plan: setting it with no effect would give the illusion of an active protection. When the subscription ends, its benefits stop without losing the configuration. Plans are available from the bot's Discord store.

Multi-server network

Link several servers you own into one network (Premium plan). A main server carries the subscription and the configuration; linked servers need nothing. A network has at most 9 linked servers besides the main one.

• Shared bans: banning someone on one server bans them everywhere in the network. Only bans travel, and only if they succeeded at the source.
• Forwarded logs: logs from linked servers arrive in a channel of the main server, by webhook, each under the name of its source server.
• Shared channels: a message posted on one side appears on the other with its author's name and avatar. Each channel sends, receives, or both.

No mention crosses a bridge: roles, @everyone and @here are neutralized in the text and refused by the API. Files are not copied, only their name is shown. Why only your own servers? Linking a server gives the power to ban there: the restriction guarantees the network gives you no power you did not already have. The condition is checked again on every ban.

Outgoing webhook

With Premium Plus, Protector sends every security incident to your own HTTPS server, to archive it or trigger your alerts. Every request is signed: you can prove it comes from Protector and was not altered.

Setup

• /webhook setSets the address and shows the signing key (owner only).
• /webhook keyShows the key again, privately.
• /webhook rotateGenerates a new key if it leaked; the old one stops working at once.
• /webhook testSends a test event.
• /webhook statusState, last delivery, consecutive failures.
• /webhook disableStops sending without deleting the configuration.

The address must use HTTPS, and local or private addresses are refused (localhost, 127.x, 10.x, 192.168.x, 172.16.x to 172.31.x, 169.254.x, .local). After 20 consecutive failures, delivery turns itself off.

What Protector sends

A POST request with a JSON body and these headers:

{
  "type": "incident.nuke",
  "guild": { "id": "123456789012345678", "name": "My server", "member_count": 1250 },
  "culprit_id": "987654321098765432",
  "adder_id": null,
  "description": "Mass destructive actions detected",
  "guild_id": "123456789012345678",
  "sent_at": "2026-07-31T02:14:07.512Z"
}

Verifying the signature

The signature is an HMAC-SHA256 of the string timestamp + "." + raw body, computed with your key. Use the raw body, before any JSON parsing: re-encoding it would change the bytes.

import express from 'express';
import { createHmac, timingSafeEqual } from 'node:crypto';

const SECRET = process.env.PROTECTOR_SECRET; // the key shown by /webhook set
const app = express();

app.post('/protector', express.raw({ type: 'application/json' }), (req, res) => {
  const ts = req.get('X-Protector-Timestamp');
  const sig = (req.get('X-Protector-Signature') || '').replace('sha256=', '');
  const raw = req.body.toString('utf8');

  // 1) Refuse requests that are too old (replay)
  if (Math.abs(Date.now() / 1000 - Number(ts)) > 300) return res.sendStatus(408);

  // 2) Compare in constant time
  const expected = createHmac('sha256', SECRET).update(`${ts}.${raw}`).digest('hex');
  const a = Buffer.from(expected), b = Buffer.from(sig);
  if (a.length !== b.length || !timingSafeEqual(a, b)) return res.sendStatus(401);

  const event = JSON.parse(raw);
  console.log('Incident received:', event.type, event.guild.name);
  res.sendStatus(200);
});

app.listen(3000);

import hmac, hashlib, time
from flask import Flask, request

SECRET = "your_key"
app = Flask(__name__)

@app.post("/protector")
def protector():
    ts = request.headers.get("X-Protector-Timestamp", "0")
    sig = request.headers.get("X-Protector-Signature", "").replace("sha256=", "")
    raw = request.get_data(as_text=True)
    if abs(time.time() - int(ts)) > 300:
        return "", 408
    expected = hmac.new(SECRET.encode(), f"{ts}.{raw}".encode(), hashlib.sha256).hexdigest()
    if not hmac.compare_digest(expected, sig):
        return "", 401
    print("Incident received:", request.json["type"])
    return "", 200

Good practices

• Answer quickly with a 2xx code: beyond 8 seconds, the delivery counts as a failure.
• Refuse requests older than 5 minutes to prevent replays.
• Compare signatures in constant time, never with ==.
• Never publish your key. When in doubt: /webhook rotate.

Community tools

• Welcome: welcome message with an optional image card and auto role (disabled when verification already handles joins). Message variables: {user} (member mention), {server} (server name) and {count} (member count). Common variants such as {mention} or {server name} are understood too.
• Tickets: private channels between a member and the team, claiming, internal notes, closing with a transcript given to the requester, automatic closing of inactive tickets and satisfaction rating. /ticket panel posts the panel. From the website, the panel can be customized: reasons offered, color, button text and opening message.
• Suggestions: /suggest posts an idea with votes, a progress bar and a discussion thread; the staff accepts or refuses with a reason.
• Giveaways: /giveaway with automatic draw. One active giveaway at a time for free, unlimited with Premium.
• Button roles: /roles add creates buttons so members can give themselves roles, 25 per channel.
• Counters: /counters creates channels showing the number of members, humans or bots, updated every 10 minutes.
• Temporary roles (Premium): /temprole gives a role that removes itself when it expires.
• Invites: /invites shows who invited whom. When Discord gives no way to know who invited a member (Server Discovery, an app that adds members, a link created while Protector was offline), the log says so instead of making up a name.
• Embeds: from the website, write rich messages with a faithful Discord preview, start from a template, and schedule or repeat them.
• Levels: XP for chatting (one gain per period, never for a message stopped by the filters) and, optionally, in voice. /level shows a member's card (avatar, username, level, progress bar, XP). Level-up announcements in the message channel, a chosen channel or by direct message, with a custom text: {user} (mention), {username} (username), {level} (level reached), {xp} (total XP), {rank} (leaderboard position), {server} (server name). Reward roles per level, curve, multipliers and channels or roles without XP are set on the site. The card background is Premium and goes through the anti-NSFW filter.
• Temporary voice channels: joining a creator channel opens a voice channel named after the member, deleted as soon as it is empty. Its owner gets a panel: open, closed or private, whitelist and blacklist, purge, mic, video, soundboards, status, transfer, settings (name, slots) and save for their next channel. Channels are stored in the database: after a restart, Protector finds them again, deletes those emptied in the meantime and serves members still waiting in a creator channel. Creator channels: 1 free, 10 with Premium, unlimited with Premium Plus.
• Notifications: new YouTube and TikTok videos and Twitch streams, announced in the chosen channel with a role to mention and a custom message ({name}, {title}, {link}, {platform}). 1 followed account free, all of them with Premium. Following an account does not announce its old videos.
• Auto responses: a trigger (contains, whole word, exact message or start of the message), a response, in every channel or just one. Case and accents are ignored, a response never mentions @everyone or a role, and messages stopped by the filters trigger nothing. 5 responses free, 25 with Premium, 50 with Premium Plus.

Permissions and diagnostic

A protection can be enabled and still do nothing if the permission it needs is missing. The worst case: anti-nuke without View Audit Log. Discord then sends no audit events, and nothing is detected.

/diagnostic shows, module by module, what works, what is degraded and what is out of service, naming the exact missing permission. Role hierarchy is checked too: a role placed too low makes "Manage Roles" useless. If a protection is out of service, the owner is warned privately, only once while the situation does not change.

During an attack, if a countermeasure fails for lack of permission (kicking raid accounts, neutralizing a nuker, restoring the server…), the owner immediately gets a private message saying what to fix.

Bot status. Protector is split into several "shards", each handling part of the servers. The shard status page shows whether each one responds, and your server's page says which shard it depends on: if a protection seems silent, that is the first place to look.

User app

Protector can be installed on your account, not only on a server. It then follows you into your DMs, group chats and servers where the bot is not present. An app installed on a profile receives no server events: these are personal tools, and they do not protect a server.

• /scan: checks a link or an image before you click, with the same scam domain list and image model as on protected servers. Only you see the answer; ten seconds between two scans, thirty with an image. For a link, the page is also opened in an isolated browser on Protector's servers, cut off from their internal network, which attaches a screenshot: you see where the link leads without opening it yourself.
• Report to Protector (right-click a message, Apps): sends a scam to the team. Ten reports per day, each read by a human.

Command list

• /autoconfigFull automatic setup.
• /configInteractive configuration panel.
• /informationsInteractive guide to every system.
• /diagnosticHealth check of the installation.
• /verification-setupVerification gate.
• /warn /mute /kick /banModeration sanctions.
• /tempban /unmute /unbanTemporary ban and lifting sanctions.
• /historyA member's record.
• /clear /slowmodeBulk delete, slowmode.
• /lockdown /unlock /panicEmergency measures.
• /backup /restoreServer backup and restore.
• /wordfilterBanned words list.
• /invites /statsInvites, moderation statistics.
• /info-server /info-user /avatarInformation cards.
• /ticket /suggest /giveawayTickets, suggestions, giveaways.
• /roles /counters /temproleButton roles, counters, temporary roles.
• /audit /logs /webhookAudit, persistent log and outgoing webhook (Premium Plus).
• /levelA member's level card.
• /scanCheck a link or an image.
• /appealAppeal a blacklist.
• /support /voteSupport server invite, vote on top.gg.
• Report…Right-click menus: report message, report user, or report to Protector.

Languages

The bot and this website are available in English, French, Spanish and German. On Discord, the choice is made in /config, Language, and applies to everything the bot writes on the server. A sentence not yet translated shows in English rather than as raw text. On the website, the selector at the top of the page remembers your choice.

FAQ

Is Protector free?

Yes. All protection, moderation and management features are free. Paid plans only add optional analysis.

Anti-nuke says "action impossible", why?

Protector's role is not high enough, or a permission is missing. Move it to the top of the role list and run /diagnostic.

Does restoring bring messages back?

It rebuilds the recent text messages of each channel through a webhook, if the option is enabled in /config, Anti-nuke. Other content is not kept.

Why is a module greyed out on the dashboard?

Either it is protected and you are neither the owner nor a member of the security role, or it requires a Premium plan the server does not have.

Is my data safe?

Protector only stores what it needs and analyzes messages in real time without keeping them, except for the cases described in the privacy policy.

I did not find my answer.

Join the support server: the team answers questions and feedback.