Discord moderation that shows its work.
An anti-nuke that reverses what an attacker did instead of alerting after the fact. Automod filters that compose predictably. Delegated moderator authority without handing out Discord permissions.
What the bot is reporting.
Read from the snapshot the bot publishes to Redis. A figure it could not read renders as a placeholder, never as a zero.
Built for the day it goes wrong.
Open any module for what it does and who can configure it.
Three steps to a configured server.
Invite the bot
Add Karen with the invite link and accept the permission set it asks for. Anti-nuke cannot reverse a change it has no permission to reverse.
Sign in with Discord
The dashboard lists the servers you own or administer, as Discord reports them, and opens on the ones where Karen is installed.
Configure each module
Every module has its own console. Anti-nuke, panic and warden answer only to the server owner and extra owners; the rest to administrators.
Who can change what.
Each gate is enforced by the API, not by hiding a button. Being able to open the dashboard is not the same as being able to arm it.
The access model in fullServer owner or extra owner only
Anti-nuke, panic, and warden. A server administrator is not enough, and there is deliberately no bot-owner bypass — the modules that can undo a mass deletion are also the ones an attacker with Administrator would disable first. Extra owners are added by the server owner.
Administrator or Manage Server
Everything else in the dashboard: automod, logging, autorole, tickets, containers, and the image block list. Ownership of the server also satisfies this.
Delegated, without the Discord permission
Warden grants one member one moderation capability with its own rate cap, so a moderator can ban without being handed Ban Members — which would also let them ban without limit and outside the bot’s audit trail.
Add Karen to a server.
Invite the bot, then sign in with Discord to configure it from the dashboard.