SYSTEM: SECURED // SHIELD ACTIVE
Initializing ContextGuards
Gemini · Structured output · Self-hosted

Clean chat
without manual
moderation.

Next-generation AI moderation for Telegram groups and channels. While competitors manually clean spam, you focus on community growth. Deploys to your Cloudflare in five minutes.

<2sec
AI reaction
24/7
without pauses
$0
hosting · free tier
AI decision 47ms
{
  "action": "delete",
  "category": "spam",
  "confidence": 0.99
}
Today
847 attacks blocked
Trusted by community owners
Crypto Research EN
IT Broadcast
Startup Digest
Dev Channel
Product Weekly
Problem

Same thing every day.
While you sleep, your chat gets overrun.

01 · Bots

Crypto scams, casinos, "easy income"

Dozens of identical messages per hour. Links to fake exchanges, schemes, Pump & Dump.

73% of new members · bots
02 · Scammers

Fake admins,
phishing, impersonation

DMs "from the admin", fake airdrop links. Members lose money and trust.

€2.8B stolen · Telegram 2025
03 · Toxicity

Insults,
hate speech, spamming

Active members leave. Manual moderation does not scale. The community slowly dies.

churn · without AI
Features

Not a bot.
An AI platform.

Gemini analyzes context, intent, and patterns. Structured output returns the decision, category, and model confidence for every message.

Zero False Positives
Gemini deeply understands context and sarcasm. The bot does not ban regular users by mistake. 0 false positives.
// POST /v1/moderate
// latency: 47ms

input: "Airdrop 1000$ → t.me/fake_exchange",
decision: {
  action: "delete",
  category: "crypto_scam",
  severity: 0.97,
  ban: true
}
Groups + Channels
The only solution with full moderation of comments in channels.
Autonomy & Failover
Your hosting. Built-in Failover: if Gemini goes down, traffic instantly switches to OpenRouter. 100% uptime.
Any language
LLM natively works with all languages — without dictionaries or training.
RU
EN
ES
ZH
AR
DE
HI
+48
Native Mini App
Forget clunky websites. Manage moderation, view analytics, and approve edge cases right inside Telegram.
Reaction < 200 ms
47ms
p50 Gemini · messages are deleted before anyone sees them.
Interface

ContextGuards
in Detail.

Take a look at the interface and logic of our AI moderator. High-tech design with perfect spam detection accuracy.

Incoming Chat
Dmitry Hey guys, when is the mainnet launching?
IDLE
{
"verdict": "SPAM",
"confidence": 0.998,
"reason": "airdrop_phishing",
"action": "delete_and_mute"
}
Real-time moderation

Intelligent
context analysis

Every incoming message is scanned instantly. The bot doesn't just look for stop words; it understands the user's true intent, recognizing even hidden crypto-spam, ads, and phishing links.

Instant verdict: Scanning takes just 47ms thanks to the Gemini 2.0 Flash model.
Structured Output: The model returns strictly valid JSON with spam categories and confidence levels.
Telegram Mini App

One-click management
right in your chat

Forget about complex external websites and control panels. A convenient native Telegram Mini App gives access to complete analytics and configuration right inside the messenger.

Everything at hand: Track deletion counts and view live log history.
Quick settings: Enable Shadow Mode, Night Mode, or Anti-Raid protection in one tap with haptic feedback.
Moderator
RU EN ES
Log out
All
Active
Muted
Banned
VIP
Sort:
Violations
Activity
Quality
T
36 2
S
26 1
A
16 75
0
Total (7d)
0
AI errors
0
Download
No errors
Bot is running smoothly
Content Moderation
Text Check text messages
Voice Transcribe and check voice
Images Analyze photos and media
Action Thresholds
Delete Points before deletion
-
40
+
Mute Points before restriction
-
70
+
Customization

Full control
over AI.

No coding required. Manage complex logic via the Telegram Mini App. Toggle modules on and off with one click.

Voice, Photo & Video
Gemini Vision reads text on images on the fly, while audio is transcribed to analyze for spam and scams.
Night Mode
Set stricter ban and deletion thresholds for nighttime. The bot guards the chat while you sleep.
AI Appeals Killer Feature
A banned user can submit an appeal. AI analyzes the context and can lift the ban automatically.
Anti-Raid & Shared Blocklist
During a massive bot attack, the chat enters Lockdown mode. Spammers from your other chats are banned preemptively via a Shared Blocklist.
Shadow Mode
Afraid to trust the neural network with deletions? Enable Shadow Mode: the bot only writes to the admin panel who it would have blocked, without physically deleting anything. Build trust with AI safely.
Installation

One command.
Five minutes.

The script automatically deploys Workers, D1, KV, registers the webhook, and spins up the Mini App. If Node.js is missing, it installs it.

1RUN

Run the script

One terminal command. Checks OS, permissions, and dependencies.

curl -fsSL ai-mod.sh | bash
2AUTH

Two keys

Bot token from @BotFather and Gemini API key. Both are validated automatically before deployment.

validate · deploy
3LIVE

Add to chat

Promote the bot to admin with message deletion rights. AI starts working instantly.

✓ protecting 1 chat
Comparison

How it differs
from the rest.

Feature
ContextGuards
Combot
Shieldy / Rose
SaaS AI Bots
AI context analysis
Gemini
No (Regex)
No
Zero False Positives
Yes
Many false bans
Many false bans
Depends on prompt
Channel moderation
Partially
No
No
High Availability (Failover)
OpenRouter
Crash on API fails
Data protection (Self-hosted)
You own it
Centralised processing
Third-party servers
Data on their servers
Multilingual without dictionaries
Any language
Swear lists only
Swear lists only
Management & Analytics
Telegram Mini App
Web Dashboard
Console commands
Web Dashboard
Hosting cost / Plan
$0 / Cloudflare
From $5/mo per group
Free
From $20 to $99/mo
Testimonials

Who already
sleeps at night.

"I used to spend an hour a day cleaning spam. Now the bot catches 99% itself. In a month — zero false positives."

A
Andrew K.
Crypto · 12K

"The key is AI understands context. Old bots banned people for saying 'ban' in a harmless sentence. That doesn't happen here."

M
Maria D.
IT community · 5K

"Installed in five minutes. No Docker, no servers. Runs for free on Cloudflare — checked the billing, it's $0."

D
Dmitry L.
DevOps · 8K channel
Pricing

One payment.
Forever.

Pays for itself in 2 months. SaaS competitors charge monthly fees. Our bot on your Cloudflare costs $0/mo forever — or let us deploy it for you, done-for-you.

Launch Special
Starter
One chat · full AI power
99one-time
Regular price €199
  • 1 group or channel
  • Gemini moderation
  • Mini App dashboard
  • Basic analytics
  • Community support
Popular
Pro
Up to 5 groups or channels
249one-time
Regular price €499
  • Up to 5 groups or channels
  • Gemini + OpenRouter fallback
  • Advanced Mini App
  • Advanced analytics + export
  • Community support
  • Future updates
60% OFF
Enterprise
Unlimited groups & channels
599one-time
Regular price €999
🔥 Early Adopter Lifetime License
  • Unlimited chats
  • All Pro features
  • Community support

Premium Support & Setup

All plans above include basic Community Support. If you need turnkey setup, direct Telegram communication with developers, priority bug fixes, and help with fine-tuning rules, subscribe to our Premium Support.

Done-for-you

Managed — we host it for you

Don't want to deal with Cloudflare and API keys? We deploy and run the bot on our infrastructure. Pay a one-time setup per tier, then top up your AI-moderation balance — ≈€0.80 per 1000 messages, balance never expires.

Starter setup
€299
1 chat · done-for-you
Pro setup
€499
up to 5 chats · done-for-you
Enterprise setup
€1500
unlimited + customizations

Enterprise also includes a monthly retainer (from €99/mo) for ongoing maintenance and custom work.

🤝 Get it done-for-you
Lifetime license Free updates Telegram support
Definition

What is
ContextGuards?

ContextGuards is a self-hosted AI moderator for Telegram groups and channels, designed to automatically combat spam, scams, and toxic behavior.

Unlike classic keyword-based bots, ContextGuards uses the advanced Google Gemini neural network for a deep understanding of message context. This ensures zero false positives and reliable protection.

The bot is deployed on the serverless Cloudflare Workers architecture, guaranteeing 100% data privacy and no recurring hosting fees.

FAQ

Frequently
asked questions.

No. The bot deploys to Cloudflare Workers — the free tier covers up to 100,000 requests per day. You only need a free Cloudflare account and a Gemini API key (also free).
$0 for most chats. Cloudflare free tier: 100K requests/day. Gemini free tier: 15 requests/minute. For highly active chats (1000+ msgs/hour) — Workers Paid $5/mo.
Run the installation script again — it will detect the existing setup, update the code, and preserve all settings and data. Zero downtime.
Only your Worker on your Cloudflare. Messages are sent to the Gemini API for analysis and are not stored anywhere. We have no access to your data.
Pro includes OpenRouter fallback — automatic switching to alternative models (Llama, Mixtral) if Gemini is unavailable. You don't depend on a single provider.
Yes. It moderates comments on posts via the linked discussion group. It's the only solution on the market with full channel support.

Stop cleaning spam
manually.

Free up your time to scale your business. AI handles the moderation. Deploy in five minutes.

Instant access · lifetime license