Discord Anti-Link Bot: Block Invites, Scam Links and Self-Bots

Which links cause the trouble

Not every link is a problem, so one blunt rule does not work. Servers usually fight three different things.

1. Invite links to other servers, posted to pull your members away or to advertise.
2. Scam links that copy a real site, such as a fake Nitro gift page.
3. Self-bot waves, where a hijacked account posts the same image and link in channel after channel within seconds.

Noxyr treats them as separate rules, so you can turn on the ones you need and leave the rest.

Invite link filter

Run /security antiinvite enabled:True. Messages with an invite to another server are deleted. These are never filtered: admins, roles you allow, channels you allow, and by default invites to your own server.

A partner channel or an advertising channel is the typical allow channel.

Scam links

The scam link rule deletes links that pretend to be Discord, Steam or Nitro. It catches lookalike letters, punycode and fake gift pages, and it also blocks any domains you add. Your allowed domains are never touched, and server admins are skipped. It is part of Noxyr's protection rules, configured with /security scamlinks, and every rule starts off.

Self-bot filter

Scam self-bots post one picture and one link many times fast. Run /security selfbot enabled:True. By default, when a member posts the same picture and link 3 times inside 120 seconds, every copy is deleted and the member is timed out for 60 minutes. You can change the threshold (2 to 10), the window (10 to 600 seconds) and the timeout.

Deletes are bundled per channel into one request, so a wave across ten channels does not hit Discord's rate limits.

Block GIFs too

A related rule, /security gifs enabled:True, deletes GIF uploads, Tenor and Giphy links, links ending in .gif and GIF stickers. Admins and allowed roles can still post them, and the member sees a short note that GIFs are off.

Switch a rule off in some channels

A single memes channel should not mean GIFs are blocked everywhere. Use /security ignore:

• /security ignore action:Add rule:GIF block channel:#memes allows GIFs there.
• /security ignore action:Add rule:Every rule channel:Staff runs no rule in the Staff category.
• /security ignore action:List shows every ignored channel and category.

Choosing a category covers every channel and thread in it, including ones made later. Every rule except anti-nuke can be ignored this way.

Permissions to check

The Security module must be on, and the bot needs Manage Messages to delete and Moderate Members to time people out. Everything is also on the dashboard under Safety, then Security, with an Ignore in these channels box on each rule. Set a staff alert channel first, so you can see what each rule caught.

Choosing the right action

Deleting the message is enough for a member who pastes an invite by mistake. For a repeat offender, add a timeout. A good split for most servers:

• Invite filter: delete only at first (timeout_minutes at 0), then add a short timeout if people keep testing it.
• Self-bot filter: keep the timeout. A real person almost never posts the same picture and link three times in two minutes, and the account is probably hijacked.
• Scam links: keep it on and check your alert channel in the first days for false positives, then add your own domains to the allowed list.

Records of what a protection rule caught are kept for 90 days, then deleted.

Common mistakes

Admins and allowed roles are skipped by design, so test the filter with a normal account, not your own. If invites are not being deleted, check that the Security module is on and that the bot can manage messages in that channel. If a partner channel is being cleaned, add it with allow_channel.

What next

Pair link rules with anti-spam for message floods and anti-nuke for staff abuse. The details are in the invite and self-bot filter docs and the protection docs. If a member account was taken over, share what to do when a Discord account is hacked. Add Noxyr to turn the filters on.