🧠Trained on your chat New
Upload your chat history — Telm learns how your people talk and vetoes false alarms, even at 97% confidence.
Telm is a Telegram anti-spam bot: it screens every message through blocklists, 150+ rules and layered AI. It starts in monitoring mode, so you see every decision before it acts.
One dashboard for every group you run — watch a 30-second live tour.
Watch every decision before it happens — with network intelligence and AI that almost never misfires.
Telm starts in monitoring mode: watch every decision it would make — and switch on enforcement only when you trust it.
Banned in one Telm community — flagged in yours. CAS plus our own federated network.
Layered checks, with an LLM only for borderline cases — 0.07% false positives on real traffic.
Text, images, voice notes and QR codes — spam can't change format to sneak past.
Upload your chat history — Telm learns how your people talk and vetoes false alarms, even at 97% confidence.
Dashboard, bot replies and appeal notices — in 25+ languages.
Add the bot and watch monitoring mode work. Free, no card — with a 7-day Pro trial.
Content, engagement and revenue — built into the same bot.
Posts on schedule, straight from your RSS — with AI rewriting them into your group's voice.
Charge for access with cards or Telegram Stars — no extra bots, a transparent platform commission from 5%. On any plan, including Free.
Reach members directly — with quotas that respect them.
Welcome flows, FAQ auto-answers, leaderboards — and a weekly digest in your DM.
Scam links, fake airdrops and wallet drainers — stopped cold.
Raids and toxicity handled while you play.
Questions become tickets with SLA and AI drafts.
A safe space for students, announcements stay clean.
Professional communities without the noise.
Protect your fans — and monetize your inner circle.
Telm says hi and starts watching — nothing is deleted yet.
See exactly what it would do in the journal. Happy? Switch on enforcement.
Everything you need to know about Telm
Add Telm to your group chat as an admin. It screens every message through blocklists, 150+ rules and layered AI — known spam is removed in under a second. And it starts in monitoring mode, so you can review every decision before enforcement is on.
Yes. Bots on Telegram physically can't delete a group, demote you or other admins, or remove members in bulk — Telm acts on regular members one decision at a time, only within the permissions you grant. Every action is written to the journal, bans and mutes can be undone in a click, and until you trust it you can keep it in monitoring mode, where it doesn't act at all.
Crypto scams, phishing links, flood attacks, fake admins, spam in images (via OCR and image hashing), QR codes — and even spam in voice notes, which get transcribed and checked like text on the Pro plan and above.
False positives are rare — 0.07% on real traffic. If Telm does slip, every action can be appealed right in the group: an AI arbiter reviews the case, usually in under a minute, and confirmed mistakes teach the system. You can also train a personal model on your group's chat export so Telm learns how your community talks.
Layered detection (blocklists, rules, our own ML, an LLM for borderline cases), fair enforcement with AI-reviewed appeals, and growth tools — including paid group access — in one bot.
Yes. One account manages any number of groups. Free covers up to 5 groups, Basic and Pro focus on one group each, Business covers 3 — and organizations scale to 50 groups on one plan.