Skip to main content
This guide walks you through setting up your first AI trader on Hyperoru.

​
Step 1: Create your account

Sign up on the Hyperoru dashboard to access the platform. Once logged in, you’ll see the main dashboard with your trading overview.

​
Step 2: Set up an AI trader

1

Create a trader

Navigate to AI Traders and click Create new trader. Give it a name (e.g., “BTC Trend Trader”) and select your preferred LLM model.
2

Configure the LLM

Enter your API key for the LLM provider (OpenAI, Anthropic, Google, DeepSeek, or any OpenAI-compatible endpoint). Choose a model like gpt-4o or claude-sonnet-4-20250514.
3

Connect your wallet

Go to the trader’s settings and bind your exchange wallet:
  • Hyperliquid: Create an API Wallet on the Hyperliquid website, then paste the agent private key and master wallet address
  • Binance: Paste your API key and secret key from the Binance Futures dashboard
Start on testnet to practice without risking real funds. You can switch to mainnet later from the trading mode settings.

​
Step 3: Create a strategy

Choose between two strategy types:

Prompt strategy

Write natural language instructions that an LLM interprets to make trading decisions. Best for complex judgment calls.

Program strategy

Write Python code that executes deterministic trading logic. Best for precise rules and backtestable strategies.

​
Step 4: Set up signals

Signals define when your strategy should analyze the market. Navigate to Signals and create a signal pool with:
  • The symbols you want to monitor (e.g., BTC, ETH, SOL)
  • Trigger conditions (e.g., OI surge, funding rate spike, price breakout)
  • Trigger logic (AND/OR between conditions)

​
Step 5: Activate and monitor

  1. Bind your strategy (prompt or program) to the trader
  2. Configure the signal pool as the trigger source
  3. Toggle Start Trading to activate
  4. Monitor performance in the Arena view or Analytics dashboard
Trading cryptocurrency perpetual contracts involves significant risk. Always start with small positions and use testnet for initial testing.