Moderation settings

Every setting below lives in the Limzo Mini App:

1. Open the Mini App: tap Open Admin Panel in your chat with Limzo, or send /settings in the group, and pick your group.
2. In the Protection section of the menu, switch on Moderation. It starts off for every new group.
3. Open Spam & links and check the switches. Most protection is already on by default.
4. Open Warnings & penalties and pick the Offender action (Mute, Ban, or Detect only) and the Mute minutes (480 by default).
5. Add your own links, usernames and trusted members under Allow & block lists.

Changes save on their own.

Send /moderation in the group: it shows whether moderation is on, the offender action, and how many trusted users and blocked terms you have. To see a rule act, post from a second account that is not an admin or trusted, for example a link to a domain you added to Blocked domains. Every removal, mute and ban is listed in Mini App → Activity log, with Unban and Unmute buttons to undo one.

One switch, at the top of the Protection section, turns the rules in this guide on or off for the whole group. While it is off, none of them delete or mute anyone; their pages grey out but keep their values, so switching moderation back on restores your exact setup. Four features keep their own switches and are not covered by it: raid protection, quiet hours, the word filter, and member warnings (/warn). Lists stay editable either way. While it is off, Limzo still counts the spam it would have removed, and if a real wave gets through, it tells an admin once how much it saw — privately when it can, otherwise with a tag in the group.

The Spam & links page holds the everyday spam protection. All of it is free on every plan:

• Delete unknown links — off by default. Removes a regular member’s link to any site not on your allowed domains, wherever the link hides: typed, behind text, in a caption, a button, a poll, or an edit. Mainstream platforms in everyday conversation, such as YouTube, and government and university sites are left alone unless you block them. Full walkthrough: the link remover guide.
• Block promo @mentions — on. Removes suspicious @channel and @bot promos; normal user tags pass.
• Delete suspicious button URLs — on.
• Network shield — on. Blocks spam links and usernames already confirmed in several other Limzo groups. Your allow lists always win.
• Flood & duplicate control — on. Removes the same message pasted again and again, or a rapid burst from one member, and gives the sender a short cool-down (a brief mute, never a ban). If your admins keep undoing these cool-downs, Limzo stops removing plain repeated text in that group, tells an admin, and one tap switches it back.
• Admin alerts — off by default. Sends connected admins a private note about removals.

Nothing is posted in the group after a removal; the sender gets your group’s usual offender action.

On by default. Links that pass the other rules are checked against a continuously updated database of known phishing and malware sites — the same kind of protection browsers use. When a link is flagged, Limzo replies to it with a public warning so nobody taps it: nothing is deleted and nobody is punished, which makes a rare false alarm harmless. Warnings for the same site are limited to one per day. Pro and Ultra can switch on Auto-delete dangerous links instead. Only the link itself is ever checked — never the message, the sender, or your group.

One setting on Spam & links for everything members bring in from outside the group:

• Smart (the default) removes non-official story shares, the format spammers lean on hardest.
• Also block channel forwards adds every message forwarded from a channel or from a group posting under its own name, the way channel ads travel between groups.
• Block every forward adds forwards from people too.
• Allow everything turns it off.

A removal is delete-only and never counts toward repeat-offender escalation, unless the same member keeps bringing in the same kind of outside content over and over. Admins, trusted users, allowed usernames, your own channel, and posts your own members made in the group are always exempt. Full guide: blocking forwarded messages.

The Allow & block lists page holds your group’s own lists. Changes save automatically:

• Allowed domains — links your group trusts, such as your project’s site. They don’t override blocked terms inside the link.
• Allowed Telegram usernames — official channels, bots or partners members may mention. Limzo’s own usernames are protected automatically.
• Trusted users — members exempt from moderation: 3 on Free, 300 on Pro and Ultra. Add an @username or Telegram user ID, or reply /trust to one of their messages in the group, which also works for members without a username. Trust follows the person through a username change; /untrust removes it.
• Blocked URL / username terms — checked in links, buttons and @usernames only, never in normal text.
• Blocked domains and Blocked usernames — links and mentions that are always removed. A blocked domain wins over every other rule.
• Blocked sticker packs — see below.

The lists stay editable while moderation is off, so you can prepare them first.

Reply /blockpack to any sticker and its whole pack is blocked: Limzo deletes that sticker and every later one from the same pack. You can also send /blockpack with a t.me/addstickers link. Removals are delete-only — nobody is muted for a sticker — admins and trusted users are exempt, and the block holds even while moderation is switched off. Reply /unblockpack to one of its stickers to undo it; /unblockpack on its own lists what is blocked. Free on every plan.

On Warnings & penalties, choose what happens to the sender when a rule removes a message:

• Mute — the default, for the Mute minutes you set (480, or 8 hours, unless you change it).
• Ban — for groups with repeated obvious spam.
• Detect only — the message is still removed and logged, but the sender is not muted or banned.
• Repeat-offender auto-ban — on by default: a third confident spam detection within 24 hours turns a mute into a ban. Low-confidence removals, such as shared contact details, never count.

Mute and ban need Limzo’s Ban users right.

A shared strike counter for offenses only a human can judge:

• Any admin replies to a message with /warn, or sends /warn @username, with an optional reason. The member gets a public warning with a visible count, and every warning has a one-tap undo button for admins.
• At the limit you choose under Warnings before action (3 by default), Limzo mutes, kicks or bans — set it in At the warn limit — and the counter resets.
• /warns shows a member’s count, /unwarn removes the latest warning, /resetwarns clears them all.

Admins can never be warned, and old warnings expire together with your moderation history.

Admins can act directly in the chat: reply to a member’s message, or use @username, with /ban, /mute, /kick, /unban or /unmute.

• Add a duration like 30m, 2h or 1d to /mute or /ban for a temporary action. Without one, /mute uses the group’s mute length and /ban is permanent.
• /kick removes the member but lets them rejoin.
• Limzo’s reply to /ban or /mute carries an Unban or Unmute button, so a mistake is undone right in the chat. Only admins can use it.

The commands stay silent for regular members, admins can never be targeted, and every action lands in the Activity log.

Members can flag a problem without pinging anyone: replying to a message with /report privately notifies the most recently active connected admins, with the message preview and one-tap Delete / Mute / Ban buttons. The command and its short confirmation tidy themselves away, and built-in rate limits mean it can’t be used to flood the chat or the admins. Works on every plan; switch it off in Mini App → Chat commands.

Optional join check on the Newcomer check page, off by default. New members are muted the moment they join and get one message with an “I’m human” button. Tapping it unmutes them at once; anyone who doesn’t verify within about 2 minutes is removed and can rejoin to retry. Messages that slip in before the mute lands are removed, admins are exempt, and joins that arrive together share one check message, so the chat stays tidy during join floods. In a channel’s discussion group, people who only comment on the channel’s posts never join the group, so the check never applies to them; their comments stay covered by the rules above.

Off by default, with its own switch on the Raid protection page. A raid is accounts arriving together, faster than any admin can react, and Limzo watches the shape of your join traffic rather than any single member. When a wave hits, the group goes read-only, the accounts that arrived in it are held back, and connected admins get a DM with a one-tap “reopen now” button. The group reopens on its own shortly after, and the notice left in the chat turns into an all-clear and then clears itself. Free on every plan; Pro and Ultra can tune how sensitive it is, how long the group stays closed, and what happens to the wave — nothing, a mute, a kick, or a ban. Closing the group needs the Ban users right, and quiet hours and a raid can hold the group at the same time without either one reopening it early.

Pro and Ultra groups can close for the night on a daily schedule:

1. Open Quiet hours in the Protection section and switch on Scheduled quiet hours.
2. Set Closes at and Reopens at — 22:00 to 07:00 unless you change them, in the group’s timezone from Mini App → Timezone.
3. Optional: switch on Announce quiet hours.

Between those hours the group is read-only, and when it reopens your exact permission setup is restored, including any media restrictions you had. Admins can post throughout. The announcement turns into a short “we’re open again” note at the end and then clears itself. Changes apply within a few minutes; if Limzo ever loses the right it needs to reopen the group, connected admins are alerted at once and Limzo keeps retrying.

List words or short phrases on the Word filter page and Limzo deletes any message that contains one as a whole word — “art” never matches “start” — in new messages, edits and captions alike.

• Free: up to 5 words, deleted.
• Pro and Ultra: up to 25 words, and a choice of Delete + warn (counts on your /warn ladder), Delete + mute for 10 minutes, 1 hour, 1 day or 1 week, Delete + ban, or Leave it, digest only.
• Keyword digest (Pro): each admin can switch on Send me the digest for a private summary at most once an hour, only when something matched, with example messages and links.

Admins and trusted users are never filtered, and the filter keeps working while moderation is off. Unlike Blocked URL / username terms, it reads the text of the message itself. Tip: start with Delete, because a short word can also turn up in a harmless phrase. Full guide: blocking words in a Telegram group.

Optional and off by default: on the Language lock page, pick the language your group runs in, with an extra Also allow English switch. Messages confidently written in another language are removed and earn a regular warning on the same fair warn ladder as everything else, never an instant kick. Short messages, names, emoji, links, mixed text and romanized chat stay untouched, and admins and trusted users are exempt. Full guide: the Telegram language lock.

On by default, on the Impersonators page. It spots members who try to mimic one of your admins or Limzo itself. Connected admins get a private alert with one-tap Ban / Looks-fine buttons; choosing “Looks fine” whitelists that member for good. Pro and Ultra can also set an automatic action — mute for 24 hours or ban — that fires only on clear impersonations, never on uncertain ones. How the scam works: the fake admin scam guide.

Telegram lets anyone edit their own messages for 48 hours — long enough to post a harmless photo, wait until admins have scrolled past, then swap the file for content designed to get your group reported. When an edit replaces the photo or file in an earlier message, or slips a file into a text-only one, connected admins get a silent private alert with a link to the message and one-tap Delete / Looks-fine buttons. Nothing is posted in the group and the member is not told. Caption-only and text-only edits never trigger it. If the edited message is more than a day old, Limzo deletes it right away and tells the admins; with moderation off it only alerts. On by default on every plan, switch on Spam & links.

A few protections run without any setup:

• Telegram Desktop hijack files. Telegram Desktop apps older than version 7.2.9 can be tricked into handing over the account when a member opens a link posted next to a small file. With moderation on, Limzo removes these files and applies your usual action to the sender; with moderation off, it leaves the message and posts one warning so nobody clicks the link.
• Bots called into the chat. When a member @mentions a bot that isn’t in your group and it posts spam, Limzo removes the post and gives that member your usual offender action, as if they had posted it themselves.
• Suspicious recruiter alerts. If one member keeps adding people who all turn out to be spammers, connected admins get one private heads-up naming that recruiter (from your Top recruiters data), with Ban, Mute or “Trust this recruiter” buttons. Limzo never punishes the recruiter on its own. Only deliberate adds and named invite links count, current admins are never flagged, and these alerts follow the Admin alerts switch.

Admins and trusted users are never affected.

Check these in order:

• Pages say moderation is switched off: tap Turn it on, or switch on Moderation in the menu.
• Spam stays in the chat: Limzo needs the Delete messages right. If it deletes but can’t mute or ban, it needs Ban users too. The permissions guide shows how.
• Nothing happens to the sender: the offender action is Detect only.
• Your test message stays: you tested from an admin account, or from a trusted user. Use a second, regular account.
• A good link was removed: add its domain to Allowed domains, or the member to Trusted users, and undo the mute in the Activity log.

Still stuck? Ask in @LimzoChat or write to [email protected].