The best Discord automod bots

The distinction that decides which bot you need

“Automod bot” covers an enormous range, from a bot that blocks six banned words to one that is effectively a small programming environment with conditionals, loops and a per-member data store. Buying the wrong end of that range costs a server in one of two directions: too little rules engine and a spam pattern nobody predicted gets through every time; too much and a volunteer moderator is now debugging a scripting language for a job that needed one checkbox. The right pick depends entirely on whether a server wants configuration or code.

This is about content filtering specifically — what gets said and what happens automatically when it is. Screening who joins and stopping an insider with real permissions are different problems, covered in the security and anti-nuke rankings.

How this ranking was built, and what it is not weighted by

Every capability and price below was read from the vendor’s own site or documentation on 31 August 2026. The ranking weighs:

Rule and trigger depth — how many distinct conditions can be expressed, not just how many keywords can be blocked.
Action range — whether the engine can escalate (warn, mute, kick, ban) or only delete.
Free-tier reality — whether the rules engine itself is free, since a filter behind a paywall is not a filter a server actually has.

It is not weighted by install count, App Directory position, or marketing volume.

The comparison

Bot	Triggers / rules	Actions	Free tier	Best for
Zeppelin	34 triggers	19	Entire product free	Maximum depth — if you can get it
YAGPDB	25 automod rules + 100 custom commands	Full templating language	Free, including the compute budget	Servers willing to write logic
Circle	17 filters	Standard	All 17 free	Roblox communities, free filtering
Sapphire	12-module engine, 5 categories	Standard	Entire product free	Deep moderation with no scripting
Vetox	8 rule types	4 action types	Free, all 8×4, with exclusions	Automod plus insider-threat protection in one place
GiselleBot	10 triggers	5, explicit precedence	Free — logging is the paywall	QR-code detection inside images
Gaius	Heat/strike engine, 13 bypass codes	Configurable	Free	Usage-permission granularity
MEE6	Keyword + spam filters	Standard	Paid — crossed out on MEE6’s own free-tier table	Only if already paying for MEE6
Discord AutoMod	5 trigger types, 6 keyword rules max	4	Free, native, permanent	The floor every third-party bot builds above

Zeppelin — rule depth as the whole product, and the hardest to get

Zeppelin’s automod is not close to contested: 34 triggers across 19 actions, plus 66 log types and what this research found to be the finest permission model in the moderation domain — a 5-level dashboard cascade with expiring assignments and per-user, per-channel and per-permission-level overrides. The entire product is free. The catch is access: Zeppelin’s own documentation says hosted-bot access is “granted on a case by case basis”, the whitelist is enforced in its own source code, and there is no invite link anywhere on the site. Self-hosting is Docker on Linux under the Elastic License. It belongs at the top of this list on capability and at the bottom on reachability. Vetox and Zeppelin compared covers what changes when the automod engine is one a server can install through a standard invite.

YAGPDB — the configurability ceiling for anyone willing to write templates

YAGPDB gives away 25 automod rules plus 100 custom commands backed by a real templating language — conditionals, loops, a per-member key-value store, and a 1,000,000-operation compute budget per execution, all free. That is the highest non-AI configurability ceiling found anywhere in this research. It asks more of the person configuring it than a rule-toggle interface does, and that is the honest trade: unmatched flexibility in exchange for learning its template syntax.

Circle and Sapphire — deep free filtering, no scripting required

Circle gives away all 17 of its content filters free — the paywall on Circle sits entirely on Roblox verification, not on automod, which is an easy detail to miss when comparing pricing pages at a glance. Sapphire runs a 12-module automod engine across 5 categories, entirely free, alongside 80+ free log types — the only paid axis on Sapphire is numeric limits and Custom Branding, never the rules engine itself.

GiselleBot and Gaius

GiselleBot ships 10 triggers and 5 actions with an explicit Ban > Mute > Kick > Warn precedence stated in its own documentation, and one trigger nothing else in this comparison has: QR-code detection inside images. Its paywall shape is unusual — all logging is paid across User, Server and Voice loggers, while the automoderator itself stays free.

Gaius runs a heat-and-strike engine with 13 documented bypass codes and gives it away free, alongside an unusually granular usage-permission model: three staff tiers plus !allow/!block/!restrict/!free commands scoped to a role, member, channel or the whole server.

Vetox — free automod, plus 20 rules automod alone can’t cover

Vetox’s automod runs 8 rule types across 4 action types, free on every tier, with exclusions and per-channel settings. What automod as a category structurally cannot do — because it inspects messages, not permissions — is stop a trusted account from deleting every channel or mass-banning the server. Vetox pairs the free automod engine with a separate 20-rule protection system, also free on Starter, covering mass channel and role deletion, permission changes, mass bans and kicks, and webhook and emoji abuse, each with its own punishment type, an undo path and per-role overrides. The two systems share one dashboard and one log stream, so a server is not maintaining a content filter in one bot and insider protection in another.

MEE6 and Discord’s own AutoMod

MEE6 is worth naming plainly here because automod is one of only two capabilities in this whole moderation domain that gets paywalled at all: its own comparison table marks Moderation with a cross on the free column, at $11.99/month for Premium.

Discord’s native AutoMod is free on every server and is a reasonable floor for a small community’s basic profanity filtering — five trigger types, three preset wordsets, four actions. Its structural ceiling is a maximum of six custom keyword rules, no escalation ladder, no case record, and it filters text only, so an image-delivered scam passes through untouched. Every bot in this comparison exists in the space above that ceiling.

Which one to install

If you want	Install
The absolute deepest rules engine, and you can get access	Zeppelin
Maximum configurability, and you’re willing to write templates	YAGPDB
Deep free filtering with a simple interface	Circle or Sapphire
Automod plus insider-threat protection in one dashboard	Vetox
Image-based scam and QR-code detection	GiselleBot
A permission model matching a large, tiered staff team	Gaius

Frequently asked questions

Is Discord’s built-in AutoMod enough on its own? For a small server whose only need is basic profanity and spam filtering, often yes. It has no escalation ladder, no case record, a six-keyword-rule cap, and reads text only — which is what drives most servers to a third-party bot once they need more than a single filter.

Does a deeper automod engine mean a better one? Not automatically. Zeppelin and YAGPDB are built around scripting depth, but Zeppelin cannot be freely installed and YAGPDB requires learning a template language. A server that just needs reliable filtering with no scripting is often better served by Circle, Sapphire or Vetox.

Is automod ever a paid feature? Rarely — of the moderation and security bots checked, only MEE6 paywalls it outright, with Security Bot charging for chat automod specifically while giving away anti-nuke, the reverse of the category norm.

Can automod stop an admin from disabling the server’s protection? No — automod filters messages, not permissions, so it cannot restrain someone who can change roles or delete channels. That is a different problem, covered by anti-nuke tools and configuration auditing.

Method

Every capability and price above was read from the vendor’s own site or documentation on 31 August 2026. Absence claims are scoped to the moderation and security bots examined in this research, not to Discord’s full moderation category.