Two meanings behind one search
The first group wants entertainment: a bot that replies rudely, roasts people in a group chat or plays along with banter. That is a toy for friends and has little to do with running a company.
The second group are admins of chats, channels and communities. They need a bot that spots a message with abusive words, deletes it, warns the author and restricts repeat offenders. Shops, training centres, delivery services and any brand with a customer community in Telegram run into exactly this.
Why profanity in a chat is a reputation issue
If your company has a public Telegram group, potential customers read it too. A chat that keeps sliding into arguments looks like a shop window nobody cleans: people hesitate to ask a question and often go to a competitor instead.
Insults aimed at staff and other members are a separate problem. Manual moderation in the evenings and on weekends is unrealistic, and a reply hours later fixes nothing. A bot reacts at once and applies the same rule to everyone, so your managers do not have to play the censor.
How a profanity filter bot works
The bot is added to the group as an administrator. Telegram lets such a bot delete messages, restrict members and ban them. Everything else comes down to settings and logic.
The hard part is not deletion but recognition. In Russian and Uzbek, swearing is easy to disguise: letters are swapped for lookalikes, dots and spaces are inserted, words are typed in Latin or Cyrillic, and languages are mixed. A plain list of banned words misses these tricks, so you need text normalisation and a dictionary covering both languages that is updated over time.
- delete the offending message
- warn the author and count violations
- temporarily mute the member
- ban repeat offenders
- keep an allow-list so harmless words are not caught
- send the admin a log of what was removed and when
Common mistakes at launch
The first is a filter that is too aggressive. If the bot reacts to word fragments, it starts deleting ordinary phrases, product names and surnames. Members get annoyed and the chat empties just as it would from abuse. Test the filter on real conversations and maintain a list of exceptions.
The second is the lack of clear rules. People should know why they were warned, so pin short chat rules and let the bot refer to them. The third is forgetting the human side: give members a way to contact an admin when a message was removed by mistake.
What about a bot that swears on purpose?
For a brand, that is risky. Rude humour may work in a closed chat among friends, but in public company communication it draws complaints and hurts the impression you make. If you want a bold, sarcastic personality, choose playful humour without profanity: it is just as memorable and carries no reputational cost.
Insulting specific people can also have legal consequences, so any such scenario should be checked with a lawyer first.
More than moderation
Once a bot is in the group, it can do useful work: greet new members, answer common questions about delivery and payment through Click or Payme, collect requests in Telegram and pass them to a manager. Then the bot is not only a bouncer but part of your sales funnel.
Decide in advance what you will measure: number of removed messages, response time to customers, requests coming from the chat. Without that, you cannot tell whether the bot is paying off.
Where to start
Write down which chat needs protection, which languages people use there and what actions the bot may take. Then build the dictionary, set escalation levels and test the bot on a small group first.
markdev builds Telegram bots for businesses in Uzbekistan, including chat moderation and lead collection. If you want to bring order to your customer community, describe your task and we will suggest a fitting setup.
FAQ
Can a Telegram bot be made to swear?
Technically yes, but for a business it is a poor idea: complaints and reputational risk outweigh any novelty. A witty tone without profanity works better.
Can a filter catch Uzbek profanity too?
Yes, if it includes dictionaries for Uzbek and Russian and accounts for Latin and Cyrillic spellings and letter substitutions. The dictionary needs regular updates.
What if the bot deletes a normal message?
Add the word to an allow-list and give members a way to reach an admin. Test the filter on real conversations before launch.
Does the bot need admin rights?
Yes. To delete messages and restrict members in a group, the bot must be an administrator with the relevant permissions.
Let’s look at your task for free
Leave your contact — in a 15-minute call we’ll show where leads get lost and what to fix first.