Skip to main content
Core Protection

Multi-Layer
Anti-Spam Protection

Comprehensive spam protection using AI, CAS integration, and behavioral analysis.

  • High detection
  • Cross-group intelligence
  • Zero config needed
Spam arrivesLayered checksRemoved before it's seen

Protection Layers

CAS Database

Millions of records

Check against millions of known spammers instantly.

AI Detection

3 levels

3-level AI system for sophisticated spam.

Behavioral Analysis

Per message

Pattern recognition for new spammer types.

Cross-Group Intelligence

Network-wide

Learn from spam patterns across all groups.

What We Catch

Crypto/investment scams
Investment fraud
Phishing links
Adult content
Fake giveaways
Bot networks
Mass DM campaigns
Impersonation
Pump & dump schemes
Malware links
Channel spam
Forward spam
Millions
Known spammers in the CAS blocklist
3 levels
Rules, fast AI screen, deep AI check
Seconds
Typical time from post to removal
Zero
Actions taken in Monitoring Mode

Automatic Actions

Instant Delete

Remove spam messages immediately.

Auto-Ban

Ban repeat offenders automatically.

New User Restrictions

Extra scrutiny for new members.

Cross-Group Sync

Share bans across your groups.

Why Telm Anti-Spam?

  • Zero Configuration - Works out of the box with sensible defaults.
  • Always Learning - AI improves continuously with new spam patterns.
  • Low False Positives - Catch spam without blocking real users.
  • 24/7 Protection - Never sleeps, never misses.

New protection layers

Newcomer contact quarantine

The classic "join, drop a contact, vanish" scheme doesn't work here. New members who immediately push handles, phone numbers, or "DM me" get stopped — with soft escalation: the message is deleted first, a mute follows only if they persist.

Polite notices instead of silent bans

When someone is punished, Telm posts a short, polite notice in their language that disappears after 60 seconds — with a private appeal link. Honest members aren't publicly shamed, and mistakes get corrected fast.

How the anti-spam layers stack up

Anti-spam is not one check but several running together. A CAS lookup matches the sender against millions of known spammers reported network-wide, so an account already flagged elsewhere is stopped on sight. In parallel the 3-level AI pipeline reads the message content, and behavioural analysis watches the shape of activity — join-then-immediately-post, identical messages across chats, forward floods — rather than any single word. Cross-group intelligence ties it together: a pattern learned in one protected group informs the others, so a campaign hitting several communities gets recognised faster after its first appearance instead of being fought from scratch each time.

The practical payoff is that spam is judged from more than one angle, which keeps false positives down — a message rarely trips every layer by accident, and a real member who shares a link is not treated like a reported bot account. It works out of the box with sensible defaults, so you are protected before configuring anything. The honest limit: brand-new spam styles with no CAS record and no prior pattern lean entirely on the AI's judgement, and truly novel attacks may need a message or two before the network catches up. Monitoring Mode lets you watch all of this before you switch on enforcement.

Frequently asked questions

What kinds of spam does Telm catch?

Crypto scams, investment fraud, phishing links, fake giveaways, adult spam, bot networks, mass-DM recruiting, impersonation, pump-and-dump schemes, and more. Detection layers cover text, links, images (via OCR and image hashing), and behavior — including the "join, drop a contact, leave" pattern.

Will it punish my legitimate members?

False positives are rare — about 0.04% on real traffic — and the system is deliberately gentle: soft escalation deletes a message before ever muting anyone, newcomers get a quarantine period instead of instant bans, and every punished member can appeal to an AI arbiter that reverses mistakes in seconds.

How fast does it react?

Spam is typically removed in under a second, before most members ever see it. The pipeline runs 24/7 and doesn't depend on a human being online.

Can punished members appeal?

Yes. Every enforcement action comes with a polite notice and a private appeal flow in 13 languages. An AI arbiter reviews each case with the group's context, approves clear mistakes automatically, and each appeal gets a trackable case number.

Where teams use this

Real community types where this feature does the heavy lifting.

Ready to Get Started?

Start protecting your community today. Free plan available.