Catch fake admins
Before they scam anyone
Scammers copy an admin's name and photo to DM your members with fake 'support'. Telm spots the impostors automatically — by name, alphabet tricks and profile photo — and shuts them down.
- Instant impostor detection
- Blocks fake-admin DM scams
- Ban, mute or notify
How impostors are detected
Name normalization
The bot strips emojis, spacing tricks and invisible characters from display names, then compares them to your real admins — so a decorated copy can't hide as staff.
Homoglyphs & alphabet swaps
Look-alike letters from Cyrillic, Greek and other scripts are folded to their Latin twins, catching names that look identical to a human but differ character-for-character.
Profile-photo matching
Admin avatars are compared with a perceptual hash, so a copied or lightly edited profile picture is flagged even when it isn't a pixel-perfect duplicate.
Admin whitelist
You define exactly who counts as staff. Anyone imitating a whitelisted admin is caught; genuine admins and approved names are never touched.
What happens on a match
Ban
Remove the impostor from the group immediately, before it can message members.
Mute
Silence the account and let a human admin review the case first, for softer handling.
Notify
Alert admins about the impersonation attempt so you decide — keeping a monitoring-only mode possible.
Why fake-admin detection matters
- Stop the classic scam - The 'an admin is DMing you' trick is the most common way members lose funds; catching the impostor in the group cuts it off at the source.
- Protect newcomers - New members can't tell a real admin from a copy — automated detection does it for them, before the first scam DM goes out.
- Keep your community trusted - One successful impersonation scam damages trust in your whole group; blocking impostors keeps your name safe.
How Telm tells a real admin from a copy
Impersonation scams follow a simple script: an account copies a real admin's display name and profile photo, joins the group, and then DMs members pretending to offer support or a claim — usually to drain a wallet or steal credentials. Telm breaks that script by treating your admin list as the source of truth and testing every other account against it. Names are normalized first: emojis, extra spacing, zero-width characters and other display tricks are stripped away so a decorated copy collapses back to the string it is imitating. Then homoglyph folding maps look-alike letters from Cyrillic, Greek and other alphabets onto their Latin equivalents, so a name that is visually identical to a real admin's but built from foreign characters is still recognized as a match.
Photos are checked too. Each admin avatar is reduced to a perceptual hash — a fingerprint that stays stable under resizing, recompression and small edits — so an impostor who re-uploads a cropped or filtered copy of the real photo is still flagged. When an account matches an admin on name or photo but is not on your whitelist, Telm acts on the rule you chose: ban it outright, mute it for a human to review, or simply notify admins and leave the decision to you. The whitelist is what keeps this precise — you declare exactly who your staff are, so genuine admins are never touched and only real impostors are caught. Because the detection happens inside the group at join time, most scams are stopped before the first fake 'admin' DM is ever sent.
Frequently asked questions
How does the bot know who my real admins are?
You maintain an admin whitelist — the definitive list of who counts as staff. Every other account is compared against it by name and profile photo. Genuine, whitelisted admins are never flagged; only accounts imitating them are. This is what keeps detection accurate and avoids false positives on your actual team.
Can it catch names that use foreign letters to look like an admin?
Yes — that is exactly what homoglyph detection is for. Look-alike characters from Cyrillic, Greek and other scripts are folded to their Latin twins, so a name that appears identical to a real admin's but is built from different characters is still recognized as an impersonation attempt, even though it looks the same to a human reader.
What if the scammer changes the profile photo slightly?
Profile photos are matched with a perceptual hash rather than an exact-file comparison. The hash is a fingerprint that survives resizing, recompression and minor edits, so a cropped, filtered or re-uploaded copy of an admin's avatar still trips the detector. It doesn't need to be a pixel-perfect duplicate.
What does the bot do when it finds an impostor?
You choose the action: ban removes the account from the group immediately, mute silences it so a human admin can review the case, and notify simply alerts admins and leaves the decision to you. That last option supports a monitoring-only setup where nothing is enforced automatically.
Does this stop scammers who DM my members directly?
It stops the source. Telegram DMs themselves are private and outside any group bot's reach, but impersonation scams almost always start with the fake 'admin' being visible in your group so members trust the name. Catching and removing the impostor inside the group — before it can present itself as staff — is what prevents the scam DM from ever looking credible.
Related features
Explore more ways Telm can help your community