Language Policy — Enforce Your Group's Language

1What language policy is

Local communities are a favourite target for foreign-language spam — Chinese or Korean broadcast campaigns dropped into a Russian-speaking chat, for example, or Arabic promo in an English support room. Language policy is the direct answer: you declare which languages belong in your group, and anything written in another writing system is handled automatically.

It is a targeted tool, deliberately separate from the general anti-spam engine. You are not describing spam here — you are describing your community's language. That makes it a precise instrument, but also a powerful one, which is why it ships off by default and should be rolled out carefully.

2How detection works

The language of a message is measured locally, by its writing system (its script), not by an AI model. That makes it instant — every message is checked the moment it arrives, with no extra latency.

You set the allowed languages as a list of ISO codes (for example en, ru, es). Leave the list empty and any language is accepted, which is the default. The bot maps each code to the script it uses and then compares it against what the message is actually written in.

3Which writing systems are recognised

Each allowed language maps to one of the writing systems the bot understands. Note that all Latin-alphabet languages share one script, so allowing English also allows every other Latin-written language (Spanish, German, French and so on) — the filter separates alphabets, not individual Latin languages:

• Latin — English, Spanish, German, French, Portuguese, Italian, Dutch, Polish, Czech, Swedish, Romanian, Indonesian, Malay, Vietnamese, Turkish and more.
• Cyrillic — Russian and Ukrainian.
• Han — Chinese characters without kana.
• Japanese — text containing kana.
• Hangul — Korean.
• Arabic, Hebrew, Thai, Greek and Indic (Devanagari, Bengali and related) — each detected as its own script.

4What happens to a foreign-language message

You choose the action for a disallowed language: delete the message, or delete it and ban the sender. The default is delete, which is the safer starting point — it removes the noise without removing the person, so an occasional mistake costs a member nothing more than one deleted line.

Pick ban only once you are confident the policy is not catching real members. It is a high-impact action, best reserved for groups under sustained foreign-language spam pressure where deletion alone is not enough.

5Built-in safeguards

The policy is designed to fail safe. When a message is short or mixes scripts, the bot does not punish it — this prevents hitting real people who drop a single foreign word, a brand name, an emoji-heavy line or a quick code snippet.

Admins, the group creator, anonymous admins and whitelisted members are never punished by language policy. And hard spam — financial fraud, pornography, gambling — is banned regardless of language, because that is global protection that does not depend on a per-group language preference.

6When to use it (and when not to)

Language policy shines in a few clear situations, and is the wrong tool in others:

• Good fit — a single-language regional community repeatedly hit by CJK or Arabic broadcast spam.
• Good fit — a strictly English (or strictly Russian) group where off-language posts are always off-topic.
• Poor fit — a genuinely multilingual community, where you would either allow every script (making the filter pointless) or risk deleting real members.
• Poor fit — as a scam blocker: scams in your own language sail straight through a language filter. Rely on the anti-spam engine for that.

7Setting it up

Open your group in the dashboard, find the language settings, add the ISO codes your group uses and choose the action. Keep Monitoring Mode on for a few days first, then review the journal to confirm no real members are being flagged before you switch to enforcement.