Automated Trading with Moltbot and Hyperliquid

How Moltbot Trading Works

Moltbot's trading module connects directly to Hyperliquid L1, enabling sub-second execution on crypto perpetuals and spot synthetics. It uses Claude Opus 4.5 for market analysis and decision-making, with hardcoded risk management rules that prevent the AI from taking excessive positions. The system is designed around the principle that "you are the CIO, Moltbot is your Chief Trader."

Setting Up Trading

To enable trading, you need a Hyperliquid account and API credentials. Configure them through the vault and enable the trading skill.

# Store Hyperliquid credentials
moltbot vault set HYPERLIQUID_API_KEY your-key
moltbot vault set HYPERLIQUID_API_SECRET your-secret

# Enable trading skill
moltbot skills install nof1-trading

# Configure in config.yaml:
skills:
  trading:
    enabled: true
    exchange: hyperliquid
    riskLimit: 0.02  # 2% max position

Pre-Built Strategies

Moltbot comes pre-loaded with system prompts and logic constraints from top Alpha Arena bots. The "Don't Catch Falling Knives" strategy ensures trades only execute when aligned with higher timeframe moving averages. Market sentiment monitoring scans high-influence X/Twitter accounts, macro events, and political developments to filter noise and identify real Alpha.

Risk Management Rules

Risk management is hardcoded, not configurable by the AI. Moltbot cannot open positions exceeding 2% of account equity. Strategies with three consecutive failures are automatically benched. The self-improvement loop records expected vs actual outcomes for each trade and dynamically adjusts strategy weights.

# Risk limits (hardcoded, cannot be overridden by AI)
# - Max position: 2% of account equity
# - Trend-following: trades must align with MA
# - Auto-bench: 3 consecutive failures
# - Human approval: required for large trades

Monitoring Trades

Moltbot sends proactive trade notifications via your configured messaging channel (Telegram recommended). Small trades execute automatically, while large trades require human approval via inline buttons. You can also query your portfolio, recent trades, and PnL at any time by messaging your bot.

# Example chat commands:
# "What's my current portfolio?"
# "Show me today's trades"
# "What's the market sentiment on ETH?"
# "Pause all trading"