Your AI Agent Can Now Trade 275 Crypto Markets Autonomously

AI agents can reason about markets, read news, analyze charts, and form views — but they can't actually trade. I built an API that fixes this.

Purple Flea Trading routes through Hyperliquid — the most liquid perpetual DEX, processing $500M+ in daily volume. 275 markets: BTC, ETH, SOL, TSLA, NVDA, GOLD, EUR/USD, and more. Up to 50x leverage. No KYC.

Here's a minimal trading agent in Python:

import requests

API_KEY = "YOUR_API_KEY"
BASE = "https://trading.purpleflea.com/v1"

def open_position(market, side, size_usd, leverage=5):
    return requests.post(f"{BASE}/order", headers={"Authorization": f"Bearer {API_KEY}"}, json={
        "market": market,
        "side": side,       # "long" or "short"
        "size_usd": size_usd,
        "leverage": leverage
    }).json()

def set_stop_loss(order_id, trigger_price):
    return requests.post(f"{BASE}/order/{order_id}/stop-loss",
        headers={"Authorization": f"Bearer {API_KEY}"},
        json={"trigger_price": trigger_price}).json()

def get_pnl(order_id):
    return requests.get(f"{BASE}/position/{order_id}",
        headers={"Authorization": f"Bearer {API_KEY}"}).json()

# Open a $100 ETH long at 5x leverage
order = open_position("ETH-USDC", "long", 100, leverage=5)
stop = set_stop_loss(order["order_id"], order["entry_price"] * 0.97)
print(f"Opened {order['order_id']}, stop at {stop['trigger_price']}")

Copy Trading

An agent can follow another agent's trades automatically:

requests.post(f"{BASE}/copy", json={"follow": "agent_id_to_copy", "allocation_pct": 50})

This opens an entire design space for agent coordination — a master agent that trades, sub-agents that copy it, referral codes routing income back to the operator.

Referral: 20% of fee markup. Hyperliquid charges 0.05% per trade; we add a small markup and share 20% of that with the referring agent.

Try it: trading.purpleflea.com