Use Scam AI directly from Discord or Telegram! No need to open a browser - just type a command and get instant AI-powered analysis.
✨ Beautiful Slash Commands
Modern Discord slash commands with auto-completion and help text.
- Type
/to see all commands - Auto-complete addresses
- Inline parameter descriptions
- Works in all servers
🎨 Rich Embeds
Beautiful color-coded embeds with all analysis data.
- 🟢 Green = Safe tokens
- 🟡 Yellow = Medium risk
- 🔴 Red = High risk
- ⛔ Dark Red = Extreme danger
⚡ Instant Results
AI-powered analysis in 10-30 seconds.
- Real-time scanning indicator
- Progress updates
- Comprehensive results
- Quick action buttons
Discord Commands
Scan a Solana token for rug pull indicators
/scan BDuWXaAcs88guisXQWvdiDGDVNn7eR7Wxw3Q5KLgpump
View scanner statistics and usage
/stats
Show help message and bot information
/help
How to Add to Your Discord Server
- Click the "Add to Discord" button below
- Select your server from the dropdown
- Authorize the bot with the required permissions
- Type
/scanto start scanning tokens!
💬 Simple Commands
Easy-to-use commands that work anywhere.
- Works in private chats
- Works in groups
- Inline mode support
- Fast response times
📱 Mobile Friendly
Perfectly formatted for mobile devices.
- Clean message formatting
- Emoji indicators
- Inline buttons
- Easy to read results
🔗 Quick Links
One-tap access to blockchain explorers.
- View on Solscan
- View on DexScreener
- Share results easily
- Copy token addresses
Telegram Commands
Start the bot and see welcome message
/start
Scan a Solana token for rug pull indicators
/scan BDuWXaAcs88guisXQWvdiDGDVNn7eR7Wxw3Q5KLgpump
View scanner statistics
/stats
Show help and available commands
/help
How to Start Using Telegram Bot
- Click the "Open in Telegram" button below
- Click "Start" in the bot chat
- Send
/scan <token_address>to analyze a token - Use
/helpfor more commands
Both Discord and Telegram bots provide the same comprehensive analysis:
📊 Risk Score (0-100)
- Overall risk assessment
- Risk level classification
- Color-coded indicators
🤖 AI Prediction
- AI Score (0-100)
- Predicted class (RUG/SAFE/SUCCESS)
- Model confidence percentage
📈 Analysis Confidence
- Data completeness score
- Reliability indicators
- Token age consideration
💰 Market Data
- Current price USD
- Market capitalization
- Liquidity information
👥 Holder Stats
- Total holder count
- Top holder percentage
- Distribution analysis
🚨 Red Flags
- Detected issues list
- Warning severity levels
- Actionable recommendations
Our AI-powered scanner includes advanced detection algorithms that identify sophisticated scam patterns:
🎯 Bundle Detection
Detects coordinated buying patterns where multiple wallets purchase simultaneously at launch. Indicates potential bundled tokens or insider trading.
- Identifies synchronized transactions
- Detects coordinated wallet behavior
- Shown inline with Fresh Wallets stat
⚠️ Sybil Attack Detection
Identifies when many fresh wallets hold top positions - a common tactic to fake distribution and avoid concentration warnings.
- Analyzes top 20 holder wallet ages
- Flags suspicious patterns (14+/20 = ALERT)
- Combined with bundle analysis
🤖 Sniper Bot Analysis
Detects automated bots that instantly purchase tokens at the moment of launch using advanced MEV techniques.
- Identifies sub-second purchases
- Detects MEV bot activity
- Warns of potential manipulation
👤 KOL Wallet Detection
Identifies known influencer wallets in holder list. Presence of KOLs can indicate either legitimacy or coordinated pump schemes.
- Tracks known influencer addresses
- Shows KOL involvement percentage
- Contextualizes token promotion
📉 Coordinated Dump Detection
Analyzes selling patterns to detect coordinated dumps where multiple wallets sell in quick succession.
- Monitors large sell transactions
- Identifies synchronized dumps
- Real-time rug pull warnings
🕐 Wallet Age Analysis
Examines the age and transaction history of holder wallets to identify newly created accounts used in scams.
- Checks wallet creation dates
- Analyzes transaction history
- Flags fresh wallet concentration
✅ No Data Storage
Bots don't store any user data, scan history, or messages. Every scan is stateless.
✅ Read-Only Access
Bots only read messages to process commands. They never modify messages or server settings.
✅ Secure API
All analysis is performed via secure API calls to our backend. No sensitive data transmitted.
Want to run your own Scam AI bot? It's open source!
# Clone the repository (bientôt disponible sur GitHub)
# git clone https://github.com/ScamAI/scam-ai-scanner.git
# cd scam-ai-scanner
# Install dependencies
pip install discord.py python-telegram-bot httpx
# Configure .env with your bot tokens
DISCORD_BOT_TOKEN=your_token_here
TELEGRAM_BOT_TOKEN=your_token_here
# Start the web server (required)
python web_app.py
# In another terminal, start Discord bot
python discord_bot.py
# Or start Telegram bot
python telegram_bot.py
📖 Full setup instructions: Coming Soon sur GitHub
How do I interpret the Risk Score vs AI Score?
Risk Score (0-100): Traditional analysis based on holder concentration, liquidity, and red flags. Lower = safer.
AI Score (0-100): Machine learning prediction trained on 1000+ historical tokens. Higher = more likely legitimate.
Both scores complement each other - consider both when making decisions!
What does "Bundle: Yes" mean?
"Bundle: Yes" indicates that multiple wallets purchased the token in a coordinated pattern at launch. This could mean:
- Insider trading: Dev and friends buying before public announcement
- Bot activity: Automated systems splitting buys across wallets
- Bundled launch: Token sold in pre-arranged bundles
Bundle detection is a red flag but not always a scam - investigate further!
What's a Sybil Attack?
A Sybil Attack is when a scammer creates many fresh wallets to hold the token, making it appear well-distributed while actually controlling most of the supply. Our scanner detects this by:
- Checking wallet ages in top 20 holders
- Flagging when 14+ out of 20 are brand new
- Combining with bundle detection analysis
How accurate is the AI prediction?
Our Random Forest and Gradient Boosting models achieve ~87% accuracy based on historical data. However, scammers constantly evolve their tactics, and very new tokens may have insufficient data. The "Analysis Confidence" score tells you how reliable the prediction is for that specific token.
What should I do if a token shows low scores?
If both Risk Score and AI Score are low:
- DO NOT invest based on FOMO
- Check the red flags list carefully
- Research the team and project independently
- Look for bundle detection and sybil warnings
- Ask yourself: "Would I be okay losing this investment?"
Remember: Low scores don't guarantee a rug, but they indicate high risk!
Why isn't my scan working?
Common issues and solutions:
- Invalid address: Make sure you're using the correct Solana token address
- Very new token: Tokens under 1 minute old may not have enough data yet
- Network issues: Solana RPC can be slow during high traffic
- Bot timeout: Scans take 10-30 seconds - please be patient!
Need help or want to connect with the community? Here are all our official channels:
🌐 Official Website
Full scanner with advanced features and real-time analysis.
→ Coming Soon📖 Documentation
Learn how to use our tools, understand scores, and interpret results.
→ Read the docs💬 Telegram Support
Get help, report issues, and connect with other users.
→ Coming Soon🐦 Twitter/X
Latest updates, announcements, and community alerts.
→ Coming Soon💻 GitHub
Open source code, contribute, or run your own instance.
→ Coming Soon⚠️ Report Scams
Found a rug pull? Help the community by reporting it!
→ Coming Soon⚠️ Important Disclaimer
Always DYOR (Do Your Own Research). These bots provide analysis tools but are NOT financial advice. Never invest based solely on bot results. The AI has 87% accuracy, meaning 13% error rate. Scammers constantly evolve their tactics. Use these bots as ONE part of your research process.