Discord Content Moderation: Setup and Key Settings

Discord content moderation combines automatic filters, channel permissions, and human review to reduce spam, harassment, and unsafe posts in a server. Before you begin, you’ll need a server where you have permission to manage its settings; the exact labels can vary slightly between desktop, web, and mobile.

What You’ll Need

A Discord account and a server you own, or a role with permission to manage the server and its moderation settings.
A clear set of rules describing what is and is not allowed, plus at least one trusted moderator for a busy community.
Access to Discord on desktop or the web for the easiest setup. Some controls are also available in the mobile app.

Step-by-Step

Write and publish server rules. Decide how you will handle spam, slurs, harassment, unwanted advertising, personal information, and links. Avoid rules that are too vague to apply consistently. Add them to a read-only rules channel, and link that channel in your welcome or onboarding flow. Explain how members can report a problem and how moderators will respond.
Open the server settings. Select your server name, then choose Server Settings. You must be the server owner or have the relevant administrative permissions. Look for Safety Setup or the safety and moderation controls; Discord may adjust navigation labels over time. If a setting is not visible, check your role permissions and try the desktop or web app.
Configure AutoMod. In the safety or moderation settings, open AutoMod. Create or enable rules for commonly unwanted content, such as commonly flagged words, suspected spam, or excessive mentions. Choose whether a match should be blocked, sent to a private moderator channel, or both. Use alerts where a human needs to judge context, and test rules with harmless examples before relying on them.
Review word and keyword filters carefully. Add terms that are genuinely relevant to your community, and consider variations that users may use to evade a basic match. A filter may catch a word inside an innocent phrase or miss altered spelling. Use exceptions sparingly and only for trusted roles or channels where there is a clear reason. Never assume a keyword list can identify context or intent perfectly.
Choose how new members can participate. Review the server’s verification and membership screening options in Safety Setup or related settings. Screening can ask new members to agree to your rules before they talk. Verification requirements can make it harder for some spam accounts to join, but can also add friction for legitimate members. Select a level that fits your community, then check the current description shown in Discord before enabling it.
Set safe defaults for channels and roles. In Server Settings, review roles and permissions. Give moderation powers only to people who need them. Keep sensitive channels private, limit who can post links or mention everyone where appropriate, and make a staff-only channel for reports and AutoMod alerts. Check channel-specific permissions too: they can override what a role appears to allow.
Assign moderator roles and agree on a process. Create or update a moderator role with only the permissions needed for the job, such as managing messages or moderating members. Avoid granting administrator access casually, since it allows broad control over the server. Agree on how staff will document incidents, when to warn or remove content, and how to escalate urgent cases.
Test the setup and explain it to members. Use a private test channel or a trusted test account to confirm that blocked content is blocked and alerts reach the right staff channel. Do not post real slurs, threats, or personal information just to test a filter. Tell members where the rules are and how to report content. Review alerts and false positives after launch, then adjust the rules.

Common Mistakes

Enabling every filter at once: A broad setup can block normal conversation. Start with the most useful rules, observe results, and expand gradually.
Using a filter as a substitute for moderators: Automated systems can miss disguised content and misunderstand harmless discussion. Route uncertain cases to staff and review reports promptly.
Giving too many people administrator access: Use a dedicated moderator role and limit sensitive permissions to trusted staff.
Forgetting channel overrides: A role may have one permission at server level and a different one in a particular channel. Check both when something behaves unexpectedly.
Failing to tell members what is filtered: Publish understandable rules and provide a way to appeal a mistaken removal. Consistent, transparent enforcement builds trust.

Tips

Keep rules short enough to understand, but specific enough that members and moderators can apply them consistently.
Use a private staff channel for alerts; do not expose moderation logs or reports to the whole server.
Review AutoMod alerts regularly. If a rule repeatedly catches harmless messages, narrow it or add a justified exception rather than ignoring every alert.
Have more than one trusted moderator for active servers, and agree on coverage so reports do not depend on one person being online.
Recheck permissions after adding bots, roles, or channels. Third-party moderation bots are optional; if you use one, review its requested permissions and configure it separately.

Troubleshooting

AutoMod is not blocking a message: Confirm the rule is enabled, the selected action is blocking rather than alert-only, and the message actually matches the rule. Check whether an allowed role or channel is exempt, and test with a simple, harmless match.

Legitimate messages are being blocked: Inspect the alert to see which rule matched. Narrow the keyword or mention rule, remove an overly broad term, or add a limited exception if appropriate. Test the change before applying more filters.

Moderators are not receiving alerts: Verify the destination channel exists and that the moderator role can view it. Check the channel’s permissions and make sure the relevant AutoMod action is set to send an alert. Ask a moderator to check their notification settings as well.

A setting is missing or cannot be changed: Confirm that you are using the correct server and have the required permission. The server owner can review your role. If the setting still does not appear, update or reload Discord, try desktop or web, and look for the current safety settings under Server Settings.

Members can still post in a restricted channel: Review that channel’s permissions for @everyone, member roles, and individual overrides. Make sure an allowed role has not been granted access unintentionally, then test using a regular member account.

Frequently Asked Questions

Is Discord AutoMod enough to moderate a server?
Not by itself. AutoMod can catch or flag some patterns, but people are needed to assess context, handle reports, and apply rules consistently.

Do I need to install a moderation bot?
No. Discord’s built-in settings can provide basic automated filtering and server controls. A bot may add features, but it is optional and requires separate setup and permission review.

Can AutoMod block every offensive message?
No filter catches everything. Users can alter spelling, use images, or communicate in context a rule cannot interpret. Combine filters with reporting and moderator review.

Can I let members appeal a blocked message?
Yes. Explain how to contact moderators, such as through a support channel or designated report process. Review appeals fairly and adjust a rule if it is producing repeated false positives.

Why can a member still post after I set a filter?
The rule may only send alerts, may not match that message, or may exempt the member’s role or channel. Check the rule’s action and exceptions, then test it again.

Should I use strict verification for every server?
Not necessarily. Stronger verification can deter some spam but may make joining less convenient. Choose based on the server’s needs and review the setting’s current effect before enabling it.

General steps for current mainstream versions; menus can differ by device and version. See the disclaimer or report an error.

More in this bed

The whole Tech How-To bed →