ProductsAboutBlogContact
3-layerSpam detection
25Languages
200+Spam patterns
<100msLive updates

What is Telm?

Telm is a complete platform for running healthy Telegram communities. It watches every message in your group and stops spam, scams, and bots automatically — combining hardcoded rules, the global CAS blocklist, AI classifiers, and image analysis — so your team doesn't have to moderate by hand around the clock.

Unlike a simple anti-spam bot, Telm pairs enforcement with insight: a real-time dashboard with analytics, activity heatmaps, and member leaderboards, plus a no-code rule builder. And every group starts in Monitoring Mode, so you can trust the bot before you ever let it act.

Multi-signal detection
Rules, the CAS blocklist, AI, OCR, and image hashing work together to catch what single-method bots miss.
Safe by default
Monitoring Mode lets you watch every decision the bot would make before a single message is touched.
Insight, not just blocking
Real-time analytics, heatmaps, and leaderboards show how your community actually behaves.
Your rules, your way
Build blacklists, rate limits, content and link filters — all without writing a line of code.

Trust it before you switch it on

Monitoring Mode is what makes Telm safe to roll out in a busy group. The bot runs its full detection pipeline but takes no action — you decide when it's ready.

1

Joins in observe mode

Add Telm to your group and it starts in Monitoring Mode automatically — nothing is deleted or banned.

2

Every message is scanned

Each message runs through the full detection pipeline. The verdict is logged, but no action is taken.

3

You review the calls

In the dashboard you see exactly what would have happened and mark each detection as spam or a false positive.

4

Flip the switch

Once you trust the accuracy, turn off Monitoring Mode and Telm starts enforcing in real time.

How Telm catches spam

Three detection layers work in sequence, escalating only the messages that need a closer look — backed by the global CAS blocklist.

01

Rule engine

200+ hardcoded patterns across 30+ languages catch link spam, excessive caps, emoji floods, and crypto or casino scams in under 10ms.

02

AI classifier

Borderline messages go to an AI model that scores intent — promotional, scam, phishing, or normal — with results cached for instant repeat detection.

03

Semantic memory

Confirmed spam is embedded into a vector index, so messages that merely look like known spam get caught even when the wording changes.

On top of all three, the global CAS blocklist bans known spammers the moment they try to join.

Key Features

AI Spam Detection

Three-layer detection: hardcoded rules, a fast local classifier, and deep AI checks for uncertain cases. Catches spam other bots miss.

CAS Anti-Spam

Instantly bans known spammers using the global Combot Anti-Spam blocklist — before they ever post in your group.

Monitoring Mode

New groups start in observe-only mode. See exactly what the bot would do — delete, mute, ban — before it touches a single message.

Real-Time Analytics

Daily message and member stats, activity heatmaps, and leaderboards by messages, XP, and reputation — updated live over WebSockets.

Custom Moderation Rules

Build your own blacklists, rate limits, content filters, and link blockers with regex support. Full control, no code required.

Multi-Group & Support Mode

Manage many groups from one dashboard, with a dedicated support mode for customer-service communities and member requests.

Analytics that help you run the community

Telm doesn't just block spam — it shows you how your community lives and grows, updated live as it happens.

  • Daily messages, active members, and spam blocked at a glance
  • Activity heatmap across every hour and day of the week
  • Leaderboards by messages, XP, and reputation
  • Per-member activity: levels, XP, active days, and streaks
  • Live updates streamed to your dashboard in under 100ms

Use Cases

  • Crypto communities fighting scams
  • Tech support and help-desk groups
  • Gaming and creator communities
  • Large multi-group networks

Frequently asked questions

Q.My group is large and busy — will Telm delete legitimate messages?

No. Telm starts in Monitoring Mode and shows you exactly what it would do without touching a single message, so you can confirm the accuracy before enabling enforcement.

Q.What exactly is Monitoring Mode?

An observe-only mode that runs the full detection pipeline and logs every verdict without acting. You review the calls in the dashboard and switch to real enforcement when you're ready.

Q.How does Telm catch spam?

Three layers: 200+ hardcoded rules, an AI classifier for borderline cases, and a semantic memory of confirmed spam — plus the global CAS blocklist that bans known spammers on join.

Q.Can I write my own rules?

Yes. Build blacklists with keywords or regex, rate limits, content filters, and link or domain blockers — all without code, applied within seconds.

Q.Which languages does Telm understand?

Detection patterns cover 30+ languages and the dashboard is available in 25, so global communities are protected and managed in their own language.

Q.How much does it cost?

Free for one group with rule-based moderation, Pro at $9 for up to five groups with AI moderation and analytics, and Enterprise for unlimited groups and history.

Ready to get started?

Explore our products or get in touch to learn more.