Connect Trading Bot or AI Agent via JSON / MCP

git clone https://github.com/ipredictsport/tennis-trading-bot.git
cd tennis-trading-bot
pip install -e .

tennis-bot setup

# Preview qualified trading edges right now without placing orders:
tennis-bot run --mode paper --dry-run

# Run a single evaluation cycle:
tennis-bot run --mode paper --once

# Continuous daemon loop (monitors and executes every 5 minutes):
tennis-bot run --mode paper

tennis-bot run --mode live

You are the "Tennis Quant & Prediction Market Advisor," powered by live quantitative models from iPredictSport.com.

Your role:
1. When users ask about upcoming ATP or WTA tennis matches, query the getTennisPredictions action to fetch the latest predictions board.
2. Clearly explain model favorites, win probabilities (e.g. "Alexander Zverev 86.8% vs Botic Van De Zandschulp 13.2%"), tournament, surface, and confidence rating.
3. When users ask for betting advice or value bets:
   - Identify contracts with positive fee-aware edges (edge_pp >= 5.0).
   - Display recommended bankroll allocation using the fee-aware quarter-Kelly formula.
   - Remind users to place resting limit orders to capture maker rebates ($0 fees).
   - Include direct trade links and mention that users can claim sign-up fee credits on Kalshi via: https://kalshi.com/r/eb2fd257-2bc9-465a-a18c-5e9a0ab4848d
4. Maintain a rigorous, objective, quantitative tone. Always remind users that past performance does not guarantee future results and to practice disciplined bankroll management.

curl -O https://ipredictsport.com/community_alerts.py

# Set webhook URL:
export DISCORD_WEBHOOK_URL="https://discord.com/api/webhooks/..."

# Test connection with sample alert:
python community_alerts.py --test

# Send currently un-alerted market edges:
python community_alerts.py --once

# Set Telegram credentials:
export TELEGRAM_BOT_TOKEN="123456789:ABCdefGhIJKlmNoPQRstUVwxyZ"
export TELEGRAM_CHAT_ID="@your_channel_or_chat_id"

# Send test alert with interactive Kalshi button:
python community_alerts.py --test

# Schedule every 30 minutes in crontab:
*/30 * * * * cd /path/to/alerts && python3 community_alerts.py --once >> alerts.log 2>&1

# Or run as continuous background daemon:
python community_alerts.py --daemon

curl -O https://ipredictsport.com/mcp_server.py

{
  "mcpServers": {
    "tennis-predict": {
      "command": "python",
      "args": ["path/to/mcp_server.py"]
    }
  }
}

curl -O https://ipredictsport.com/mcp_server.py

{
  "mcpServers": {
    "tennis-predict": {
      "command": "python",
      "args": ["path/to/mcp_server.py"]
    }
  }
}

curl -s "https://ipredictsport.com/predictions.json" | jq .

const resp = await fetch("https://ipredictsport.com/predictions.json");
const data = await resp.json();

console.log("Upcoming Matches:", data.upcoming_board);
console.log("Market Edges:", data.kalshi_evaluations);