Anti-spam

Spam and ad protection for Telegram groups

Spam in a busy chat is not only “buy my course”. It is invitations to neighbouring channels, “message me privately”, exchange services, links from accounts registered yesterday, and ads hidden inside a picture.

The kinds of spam the bot catches

Advertising messages

Unsolicited commercial ads and self-promotion in any form.

Invitations to other chats

Links to channels and chats, invite links and referral links.

Contacts and phone numbers

Contact cards and phone numbers from any country, right there in the text.

External links

Links from new members and from unknown domains; the domains you need go on the allowlist.

Covert offers

Messages with no outright ad, but with a clear intent to pull someone into DMs or onto an outside site.

Service offers

“I'm an estate agent”, “I build websites”, “I swap USDT” — where the chat is not meant for that.

Altered characters

Look-alike letters from other alphabets, zeros for letters, dots and spaces inside words, Unicode fonts.

Text on images

An ad captured as a screenshot is read off the picture and checked like ordinary text.

Why a list of banned words is not enough

A list catches exactly the spelling it contains. A spammer only has to change the spelling once and the rule stops working — while a member who happens to use a flagged word is the one who pays for it.

quick cashthe filter catches it· any filter catches this
q u i c k c a s hthe filter misses it· spaces inside the words — the list already misses it
qu1ck ca$hthe filter misses it· a digit for the “i”, a symbol for the “s”
Got an interesting offer for you, message methe filter misses it· not a single banned word in it
ModerBot unfolds the disguise back into plain text before any check at all, and then weighs what the phrase means. That is why it sees all four lines above the same way — including the last one, which contains no banned words whatsoever.

What happens after something is found

The same path for every category, and every step is visible to the admin.

1. Rule
The category is identified: ads, links, scams, flooding and so on. Categories are switched on individually.
2. Decision
Context, the age of the account and the member's history are taken into account. Group admins are always skipped.
3. Action
Whatever is set for that category: skip, warn, delete, restrict or ban.
4. Notice
A short message with the category and the reason goes to the chat. It can be shortened, turned off or set to delete itself.
5. Log
The event is saved: who, what, which category, which reason. From here the action is undone in one click.

When the chat sees thousands of messages a day

In a large group manual moderation is the first thing to break: an admin physically cannot read everything, and spammers turn up at night.

  • Checks run around the clock and identically — at three in the morning just as at noon.
  • A busy chat costs no more than a quiet one: the bot is free regardless of the number of messages and members.
  • A mass influx of accounts turns on slow mode automatically and lifts it once the influx is over.
  • The /purge command clears out every message from one member at once, if a mailing did get through.
  • The log shows which categories fire most often — it tells you exactly what to fine-tune.

Frequently asked questions

Can the bot delete a message on its own, without an admin?

Yes, if the category's action is set to “delete”. The bot needs the right to delete messages for that — it is one of the rights required when you connect it.

Can links be allowed?

Yes. The links category can be switched off entirely, or left on with the domains you need added to the allowlist — everything else keeps being deleted.

Can certain members be allowed to advertise?

Yes, that is what the allowlist is for. The rules do not touch a member on it at all; you can add one from the panel or with the /whitelist command in the chat.

What do I do about a false positive?

Open the log: it shows which category fired and with what reason. From there you can undo the action and add an exception so it does not happen again.

Does the protection work in several groups at once?

Yes. Each group is configured separately — its own categories, its own actions, its own allowlist. All groups appear in one list in the Mini App.

Spammers will come back with new accounts. What then?

New accounts are checked on arrival: fresh and obfuscated profiles are logged, the CAS spammer database cuts off known ones, and a mass influx turns on slow mode. A newcomer's first messages are checked more strictly than usual.

How many messages a day can the bot handle?

There is no limit on the number of messages. There is no charge either: the bot works for free in groups of any size.

Up and running in a couple of minutes

Add the bot to your group and grant admin rights — protection starts at once, with sensible defaults.

Add to Telegram

Read next

How AI moderation worksImage checksAll rule categories