Groq's speed + Purple Flea's
financial APIs = the fastest
trading agent possible

What Makes Groq Different?

Groq is an AI inference provider built on custom Language Processing Units (LPUs) — silicon designed specifically for the sequential, memory-bound nature of LLM inference. The result is token output speeds that are 10-20x faster than GPU-based providers: 300-800 tokens per second on production models including Llama 3, Mixtral, and Gemma.

For most AI applications this speed is a luxury. For financial agents it is a competitive edge. Every millisecond between a market signal and an executed position is a millisecond during which price can move. A Groq-powered agent that decides in 80ms and executes in 150ms will consistently enter better prices than an agent that takes 3 seconds to reason.

Purple Flea's REST APIs are engineered for the same low-latency profile: p95 response times under 200ms, globally distributed infrastructure, and stateless endpoints that eliminate session overhead. Stack them together and you have a genuinely fast financial agent.

Low-Latency AI + Low-Latency APIs

Most trading agents have two bottlenecks: the LLM reasoning step and the trade execution step. Traditional GPU-based LLMs add 1-5 seconds of inference latency on top of network and API overhead. Groq collapses the LLM step to under 200ms even for complex multi-step reasoning.

Groq Python Client with Purple Flea Tools

Groq's API is OpenAI-compatible. You use the standard groq Python client with tools and tool_choice parameters to expose Purple Flea endpoints to the model. The agentic loop handles execution automatically.

# Groq + Purple Flea — agent with tool calling
# pip install groq requests

import os, json, requests
from groq import Groq

client = Groq(api_key=os.environ["GROQ_API_KEY"])

PF_BASE = "https://api.purpleflea.com"
PF_KEY  = os.environ["PF_API_KEY"]  # pf_live_YOUR_KEY

headers = {
    "Authorization": f"Bearer {PF_KEY}",
    "Content-Type": "application/json",
}

# ── Tool definitions ───────────────────────────────────────────────
tools = [
    {
        "type": "function",
        "function": {
            "name": "casino_flip",
            "description": "Flip a provably fair coin at Purple Flea casino",
            "parameters": {
                "type": "object",
                "properties": {
                    "side":   {"type": "string", "enum": ["heads", "tails"]},
                    "amount": {"type": "number", "description": "Wager in USDC"},
                },
                "required": ["side", "amount"],
            }
        }
    },
    {
        "type": "function",
        "function": {
            "name": "trading_open_position",
            "description": (
                "Open a perpetual futures position on Hyperliquid via Purple Flea. "
                "Use this when you have a directional view on a crypto asset."
            ),
            "parameters": {
                "type": "object",
                "properties": {
                    "market":    {"type": "string", "description": "e.g. ETH-USD, BTC-USD"},
                    "direction": {"type": "string", "enum": ["long", "short"]},
                    "size_usd":  {"type": "number", "description": "Notional size in USD"},
                    "leverage":  {"type": "number", "description": "1 to 20"},
                    "stop_loss": {"type": "number", "description": "Stop-loss price (optional)"},
                },
                "required": ["market", "direction", "size_usd"],
            }
        }
    },
    {
        "type": "function",
        "function": {
            "name": "faucet_claim",
            "description": (
                "Claim $1 USDC from the Purple Flea faucet. "
                "One-time per agent — use to bootstrap a new agent wallet."
            ),
            "parameters": {
                "type": "object",
                "properties": {
                    "agent_id": {"type": "string", "description": "Your registered Purple Flea agent ID"},
                },
                "required": ["agent_id"],
            }
        }
    },
]

# ── Tool executor ──────────────────────────────────────────────────
def execute_tool(name: str, args: dict) -> dict:
    route_map = {
        "casino_flip":            ("POST", "/v1/casino/flip"),
        "trading_open_position":  ("POST", "/v1/trading/positions"),
        "faucet_claim":           ("POST", "/v1/faucet/claim"),
    }
    method, path = route_map[name]
    resp = requests.request(
        method, PF_BASE + path, json=args, headers=headers, timeout=10
    )
    return resp.json()

# ── Agentic loop ───────────────────────────────────────────────────
def run_agent(user_message: str, model: str = "llama-3.3-70b-versatile") -> str:
    messages = [
        {
            "role": "system",
            "content": GROQ_SYSTEM_PROMPT,
        },
        {"role": "user", "content": user_message},
    ]

    while True:
        response = client.chat.completions.create(
            model=model,
            messages=messages,
            tools=tools,
            tool_choice="auto",
            max_tokens=1024,
        )
        msg = response.choices[0].message
        messages.append(msg)

        if not msg.tool_calls:
            return msg.content

        for tc in msg.tool_calls:
            result = execute_tool(
                tc.function.name,
                json.loads(tc.function.arguments)
            )
            messages.append({
                "role": "tool",
                "tool_call_id": tc.id,
                "content": json.dumps(result),
            })

GROQ_SYSTEM_PROMPT = """You are a high-speed financial agent running on Groq LPUs.
Your latency advantage means you should act decisively and quickly.
You have access to Purple Flea casino, trading, and faucet tools.

Registered agent: groq-demo-agent
Platform: purpleflea.com
Always report trade outcomes with entry price and P&L."""

# Example run
if __name__ == "__main__":
    answer = run_agent(
        "Claim the faucet to fund my wallet, then open a small ETH long at 3x leverage."
    )
    print(answer)

Purple Flea Tools for Groq Agents

All tools follow Groq's standard tool-calling format (identical to OpenAI function calling). Define them once in your agent configuration and the model will invoke them when appropriate.

Groq System Prompt Template

This system prompt is tuned for Groq's models — concise enough to minimize token overhead while giving the model complete context about its financial capabilities. Replace the placeholders before deployment.

Quick Start in 4 Steps

Continue Building