AI Anti-Spam Bot for Telegram

🤖 AI-powered anti-spam bot for Telegram groups

🎯 Demo Bot • 📢 Channel • 💬 Group • 简体中文

---

⚠️ Usage Notice

If you use this project, please:

• ✅ Keep the developer information in the Bot (/start command)
• ✅ Credit the source in your project
• ✅ Do not remove copyright notices from the code

This is the most basic respect for open source authors 🙏

💡 Official Version: This is the official repository maintained by the original author. Forks may be outdated.

---

✨ Features

• 🛡️ Smart Detection: AI-powered spam detection for text, images, and stickers
• 🎯 Multi-Model Support: Choose from OpenAI, Qwen, or DeepSeek
• 📊 Flexible Strategy: Configurable detection days, message count, verification times
• 🔓 User-Friendly Management: One-click unban, admin panel
• 📢 Ad Buttons: Custom advertisement buttons on ban notifications
• ⚡ High Performance: Async processing, non-blocking
• 🔄 Verification System: Verified users skip future checks, saving API calls
• 🔁 Auto Retry: AI API calls auto-retry on failure (3 times, exponential backoff)
• 📝 Persistent Logs: Runtime logs saved to data/bot.log
• 📈 Statistics: /stats command to view detection statistics
• 🌍 Multi-Language: Support for Chinese/English

🚀 Quick Start

1. Install Dependencies

pip install -r requirements.txt

2. Configuration

cp config.example.yml config.yml
# Edit config.yml with your settings

Required:

• telegram.token - Your bot token (from @BotFather)
• telegram.owners - Super admin ID (your Telegram ID, get it from @userinfobot)
• AI model config (choose one):
  • openai.api_key - OpenAI API Key
  • qwen.api_key - Qwen API Key
  • deepseek.api_key - DeepSeek API Key

3. Run

python bot.py

4. Add to Group

1. Add the bot to your group
2. Set as admin (requires delete messages & ban users permissions)
3. Send /start to test

📖 Usage Guide

User Commands

• /start - View bot info
• /admin - View admin panel (admins only)

Admin Commands

• /unban <user_id> - Unban user
• /unban (reply to message) - Unban the replied user

Super Admin Commands

• /add_ad title|link|expiry|weight - Add ad button
• /all_ad - View all ads
• /del_ad <ID> - Delete ad
• /stats - View runtime statistics (detection count, ban rate, etc.)

Ad Button Example:

/add_ad Official Channel|https://t.me/mychannel|2099-12-31 23:59:59|100

⚙️ Configuration

Basic Config

telegram:
  token: "your-bot-token"
  owners: ["your-telegram-id"]  # Super admins
  allow_any_group: true          # Allow any group
  groups: []                     # Whitelist groups (when allow_any_group=false)

# Language setting
language: "zh"  # Options: zh / en

AI Model Selection

ai_model: "deepseek"  # Options: openai / qwen / deepseek

Model Comparison:

Model	Advantage	Price	Image Support
OpenAI	High accuracy	Expensive	✅ gpt-4o-mini
Qwen	Good for Chinese	Medium	✅ qwen-vl-max
DeepSeek	Cheap	Lowest	❌

Detection Strategy

strategy:
  joined_days: 3              # Days since joining threshold
  min_messages: 3             # Minimum message count
  spam_score: 80              # Spam score threshold (0-100)
  verification_times: 1       # Verification pass limit
  check_message_count: true   # Check message count

Strategy Explanation:

1. joined_days: Users who joined more than N days ago skip detection
2. min_messages: Users with more than N messages skip detection (requires check_message_count: true)
3. verification_times: Users who passed N verifications skip detection (0=unlimited)
4. spam_score: Only ban when AI score exceeds this value

Common Scenarios:

Scenario 1: Strict Mode (Check all new users)

strategy:
  joined_days: 999999
  verification_times: 0
  check_message_count: false

Scenario 2: Relaxed Mode (Quick pass)

strategy:
  joined_days: 3
  min_messages: 3
  verification_times: 1
  check_message_count: true

Scenario 3: Balanced Mode (Recommended)

strategy:
  joined_days: 7
  min_messages: 5
  verification_times: 2
  check_message_count: true

🎨 Ad Button Feature

Ban notifications display custom buttons:

🚫 Spam Detected
...
⚠️ User has been permanently banned

[🔓 Unban]            ← Admins can click
[📢 Official Channel] ← Your ad
[💎 VIP Group]        ← Your ad

Manage Ads:

# Add
/add_ad Official Channel|https://t.me/channel|2099-12-31 23:59:59|100

# View
/all_ad

# Delete
/del_ad 1

🐳 Docker Deployment

# 1. Edit config
cp config.example.yml config.yml
vim config.yml

# 2. Start
docker compose up -d

# 3. View logs
docker compose logs -f

📊 How It Works

User sends message
    ↓
Check if admin → Yes → Allow
    ↓ No
Check if needs detection → No → Allow
    ↓ Yes
AI analyzes content
    ↓
Score < threshold → Allow + verification count +1
    ↓
Score ≥ threshold → Ban + delete message + send notification

🔧 Advanced Features

Custom Telegram API

If you have a self-hosted Telegram API server:

export TELEGRAM_API_URL="https://your-api-server.com"
python bot.py

Database Location

Default: data/bot.db

Modify:

database:
  path: "custom/path/bot.db"

📮 Contact

• 👨‍💻 Developer: 狼哥 (@luoyanglang)
• 📦 GitHub: @luoyanglang
• 🌟 Official Project: github.com/luoyanglang/AI-Anti-Spam-Bot
• 📢 Official Channel: @langgefabu
• 💬 Discussion Group: @langgepython
• 🎯 Demo Bot: @xiaolangzaibot - Add to your group to test!

💰 Support

If this project helps you, consider supporting the developer:

• ⭐ Give this project a Star on GitHub
• 📢 Share with others who might need it
• 🐛 Report bugs and suggest features
• ☕ Buy me a coffee

Your support keeps this project alive

🤝 Contributing

Issues and Pull Requests are welcome!

📄 License

MIT License

🙏 Acknowledgments

Thanks to all contributors and users for your support!

Important Notice:

• This project is open source under the MIT License
• Commercial use is allowed, but please keep the developer information
• If you modify or distribute this project, please credit the original author
• This is respect for open source spirit and motivation for continuous development

---

⭐ If this project helps you, please give it a Star!
💰 Support the developer to keep this project alive!