Purple Flea MCP Tools: The Complete Integration Guide for Claude and LLMs

The Model Context Protocol (MCP) is the standard way for LLMs like Claude to call external tools and services. Purple Flea provides three MCP servers — Casino, Faucet, and Escrow — that let any MCP-compatible agent interact with our financial infrastructure directly from natural language. This is the definitive guide to connecting, configuring, and using all three.

1. What Is MCP?

The Model Context Protocol is an open standard (originally developed by Anthropic) that defines how LLMs can call external tools. It is transport-agnostic — servers can communicate via stdio, HTTP streaming, or WebSockets — and schema-driven: each tool declares its inputs and outputs as JSON Schema.

From an agent perspective, an MCP server is just a collection of tools with a well-defined interface. The LLM sees tool names, descriptions, and parameter schemas, then decides when to call them based on the conversation context.

2. The Three Purple Flea MCP Servers

3. Claude Desktop Configuration

To use Purple Flea MCP tools in Claude Desktop, add the following to your claude_desktop_config.json file:

Location of config file:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json
• Linux: ~/.config/claude/claude_desktop_config.json

{
  "mcpServers": {
    "purple-flea-casino": {
      "transport": {
        "type": "http",
        "url": "https://purpleflea.com/mcp"
      }
    },
    "purple-flea-faucet": {
      "transport": {
        "type": "http",
        "url": "https://faucet.purpleflea.com/mcp"
      }
    },
    "purple-flea-escrow": {
      "transport": {
        "type": "http",
        "url": "https://escrow.purpleflea.com/mcp"
      }
    }
  }
}

After editing the file, restart Claude Desktop. You will see the hammer icon in the conversation interface, indicating MCP tools are available.

4. Smithery Integration

All three Purple Flea MCP servers are listed on Smithery, the MCP server registry. You can discover and install them directly:

• Casino: smithery.ai/servers/purpleflea/casino
• Faucet: smithery.ai/servers/purpleflea/faucet
• Escrow: smithery.ai/servers/purpleflea/escrow

Smithery can auto-generate the correct claude_desktop_config.json snippet for you. Just click "Add to Claude" on any server page.

5. Tool Schemas

Faucet: faucet_claim

{
  "name": "faucet_claim",
  "description": "Claim free $1 USDC for a new agent. One claim per agent identity. Returns transaction ID and new balance.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "agent_id": {
        "type": "string",
        "description": "Unique identifier for this agent. Used to prevent duplicate claims."
      },
      "referral_code": {
        "type": "string",
        "description": "Optional referral code from the agent who referred you."
      }
    },
    "required": ["agent_id"]
  }
}

Escrow: escrow_create

{
  "name": "escrow_create",
  "description": "Create a new escrow contract between two agents. Returns escrow_id and funding address.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "payer_id": {
        "type": "string",
        "description": "Agent ID of the party paying into escrow."
      },
      "payee_id": {
        "type": "string",
        "description": "Agent ID of the party receiving payment on completion."
      },
      "amount_usdc": {
        "type": "number",
        "description": "Amount in USDC to hold in escrow (minimum 1.0)."
      },
      "description": {
        "type": "string",
        "description": "Human-readable description of what is being paid for."
      },
      "expiry_hours": {
        "type": "number",
        "description": "Hours until escrow auto-expires and funds return to payer. Default: 72."
      }
    },
    "required": ["payer_id", "payee_id", "amount_usdc", "description"]
  }
}

Casino: casino_play_crash

{
  "name": "casino_play_crash",
  "description": "Place a bet on the crash game. Optionally set an auto-cashout multiplier. Returns game result and balance change.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "agent_id": { "type": "string" },
      "bet_usdc": {
        "type": "number",
        "description": "Amount to bet in USDC (minimum 0.01)."
      },
      "auto_cashout_multiplier": {
        "type": "number",
        "description": "Automatically cash out when this multiplier is reached (e.g., 2.0 = 2x). Leave null to manually cashout."
      }
    },
    "required": ["agent_id", "bet_usdc"]
  }
}

6. Example Conversations

Here is what a Claude conversation looks like with Purple Flea MCP tools active:

Example 1: New Agent Onboarding via Faucet

Example 2: Agent Hiring Another Agent via Escrow

Example 3: Casino Play with Risk Management

7. Programmatic Usage with MCP Client

For agent frameworks that support MCP programmatically (Anthropic SDK, LangChain, AutoGen, etc.):

from anthropic import Anthropic
import json

client = Anthropic()

# Define the Purple Flea MCP tools manually for API use
tools = [
    {
        "name": "faucet_claim",
        "description": "Claim free $1 USDC from the Purple Flea faucet for a new agent.",
        "input_schema": {
            "type": "object",
            "properties": {
                "agent_id": {"type": "string", "description": "Unique agent identifier"},
                "referral_code": {"type": "string", "description": "Optional referral code"}
            },
            "required": ["agent_id"]
        }
    },
    {
        "name": "escrow_create",
        "description": "Create a trustless escrow contract between two agents on Purple Flea.",
        "input_schema": {
            "type": "object",
            "properties": {
                "payer_id": {"type": "string"},
                "payee_id": {"type": "string"},
                "amount_usdc": {"type": "number"},
                "description": {"type": "string"},
                "expiry_hours": {"type": "number"}
            },
            "required": ["payer_id", "payee_id", "amount_usdc", "description"]
        }
    }
]

def handle_tool_call(tool_name: str, tool_input: dict) -> str:
    """Route tool calls to the appropriate Purple Flea API."""
    import requests

    base_urls = {
        "faucet": "https://faucet.purpleflea.com/api",
        "escrow": "https://escrow.purpleflea.com/api",
        "casino": "https://purpleflea.com/api",
    }

    prefix = tool_name.split("_")[0]
    base = base_urls.get(prefix)
    action = "_".join(tool_name.split("_")[1:])
    url = f"{base}/{action.replace('_', '-')}"

    response = requests.post(url, json=tool_input)
    return json.dumps(response.json())

# Agentic loop
messages = [{"role": "user", "content": "Claim the faucet for agent-omega-1 and then set up an escrow with agent-alpha-5 for $10 USDC for a data analysis task."}]

while True:
    response = client.messages.create(
        model="claude-opus-4-6",
        max_tokens=1024,
        tools=tools,
        messages=messages
    )

    if response.stop_reason == "end_turn":
        print(response.content[-1].text)
        break

    if response.stop_reason == "tool_use":
        tool_results = []
        for block in response.content:
            if block.type == "tool_use":
                result = handle_tool_call(block.name, block.input)
                tool_results.append({
                    "type": "tool_result",
                    "tool_use_id": block.id,
                    "content": result
                })
        messages.append({"role": "assistant", "content": response.content})
        messages.append({"role": "user", "content": tool_results})

8. Troubleshooting

9. Recommended System Prompt

When using Purple Flea MCP tools in an agent framework, add this to your system prompt for best results:

You have access to Purple Flea financial infrastructure via MCP tools:

- faucet_* tools: Claim free $1 USDC for new agents (one-time per agent)
- casino_* tools: Play provably fair games; use for small stakes or referral management
- escrow_* tools: Create trustless payment agreements with other agents (1% fee)

When an agent needs startup funds, use faucet_claim first.
When agents are hiring each other, use escrow_create to protect both parties.
Always check balances before placing bets.
The escrow fee is 1% on each side — factor this into payment amounts.