Where swearing in chats actually hurts a business
In Uzbekistan Telegram has become a working environment: residential complex chats, groups for online course students, buyer clubs, supplier and wholesale chats, and the comments under a shop channel. Wherever many people gather, rudeness eventually shows up.
An admin cannot sit in the chat around the clock. An insult posted at night or on a day off can hang there for hours, and that is exactly what new members see first. Some leave quietly, others stop asking questions. For a chat meant to warm up customers before a purchase, that is a real loss.
How a profanity filter bot works
You add the bot to the group and make it an administrator with the right to delete messages and restrict members. Without those rights it can read the text but cannot act on it. The bot checks every new message and, in a good implementation, every edited one too. Otherwise someone posts something polite and then edits it into something abusive.
What happens next depends on your rules. A typical set of actions looks like this:
- delete the message silently;
- delete it and send the author a short warning;
- count violations and mute the member for a while after a set limit;
- remove repeat offenders from the group;
- log each case in a private admin chat;
- skip messages from admins and trusted members.
The hard part: workarounds and false positives
A plain list of banned words stops working quickly. People swap Cyrillic letters for lookalike Latin ones, insert dots and spaces, stretch letters, or write Russian words in Uzbek Latin script. In local chats you also get Russian and Uzbek mixed in a single sentence, so the filter has to handle both languages.
The opposite problem is false positives. If the bot looks for a banned fragment inside any word, it starts deleting harmless phrases, product names and surnames. A customer whose ordinary question about delivery gets removed is more annoyed than by someone else's rudeness. That is why you need whole-word matching, an exceptions list and a quick way to restore a message by hand.
Ready-made bot or a custom one
For a small group a ready-made bot is often enough: connect it, pick a strictness level, done. The downsides are familiar to anyone who has tried: the dictionary may not know Uzbek expressions, the rules cannot be tuned to your niche, and data about your community lives on someone else's service.
A custom bot makes sense when you run several chats, need your own penalty logic, or want the bot to do more than moderate. It can greet newcomers, block link spam, collect leads, answer frequent questions and hand tricky cases to a manager. In that setup the profanity filter becomes one module inside a broader business bot.
Rules to decide before launch
Before launching, it helps to settle human questions rather than technical ones: what counts as a violation in your community, and how do you want to handle it.
- What is banned: only profanity, or also insults, ads and threats.
- How many warnings a member gets before a mute.
- How someone can appeal the bot's decision.
- Who is exempt: admins, managers, moderators.
- How the bot behaves in channel comments, which live in a linked group.
- Who reviews the deletion log each week and updates the dictionary.
How to tell the bot is helping
Look at practical signs: fewer complaints from members, less manual cleanup by admins, and no growth in questions like “why was my message deleted?”. The action log shows which words the bot catches most often and where it makes mistakes. If wrongful deletions are frequent, loosen the rules rather than switching the bot off entirely.
Where to start
First describe your community: how many chats, which languages people write in, what counts as rude, and who will moderate. Then decide whether a ready-made tool is enough or you need development. The markdev team builds Telegram bots for businesses in Uzbekistan and can help you make that call. If you want to discuss the task for your chat or channel, send us a request and we will go through it together.
FAQ
Does a profanity filter bot delete messages by itself?
Yes, provided it is an administrator with permission to delete messages. Without that right it can only read the chat and cannot intervene.
Will the bot catch Uzbek swearing and transliterated text?
Only if its dictionary and rules account for it. Ready-made bots are often built around Russian or English, so for bilingual chats the word list and text normalization should be configured separately.
What should I do about false positives?
Match whole words, keep an exceptions list and review the deletion log regularly. That shows you where the rules are too strict so you can adjust them.
Does it work in channel comments?
Comments live in the linked discussion group, so you add the bot there as an administrator. It then moderates them like any other chat.
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.