Warden runs its control panel inside the bot itself. Flip a setting and the very next message is judged by the new rules — no restart, no sync delay, no waiting.
Free · Self-hosted · Requires Manage Server to configure
Each one runs independently, with its own exemptions and escalation. Turn off what you don't need.
Invite links, domain allowlists, custom word filters, mention spam, excessive caps and a sliding-window duplicate detector. Edited messages get re-scanned — editing a clean message into a scam is a classic trick.
Known scam domains plus your own blocklist, look-alike detection by edit distance to discord.com and steamcommunity.com, and brand-bait heuristics. Runs on everyone, mods included — compromised accounts are the usual phishing vector.
Join-rate detection trips raid mode automatically, with an account-age gate and a manual panic switch
on both /panic and the dashboard.
Every action gets a numbered case and a mod-log embed. Separate channels for mod actions, message edits and deletes, joins, leaves, bans, and channel or role changes.
Most bots write settings to a database and hope the process notices. Warden's dashboard runs inside the bot process and writes straight into its memory.
OAuth2 with identify and guilds. Warden reads which servers you're in and
shows only the ones where you hold Manage Server, Administrator or ownership.
Servers Warden hasn't joined yet show an invite link instead. Permissions are re-checked on every request, not just at login.
The panel PATCHes a partial config to the bot, which merges it into the live cache. The next message or join event is evaluated against it immediately.
Every moderation command runs hierarchy checks, offers optional DM notification, and files a numbered case.
Add Warden, sign in, and configure it in about two minutes.