Live odds for AI agents, with nothing to install

Claude Code

One command, from the project directory. Add -s user to use it in every project.

claude mcp add --transport http puntersedge https://api.puntersedge.online/mcp \
  --header "X-API-Key: YOUR_KEY"

Claude API (MCP connector)

Claude connects to the server itself during the request; your code holds no tool logic.

import os

import anthropic

client = anthropic.Anthropic()
response = client.beta.messages.create(
    model="claude-opus-5-5",
    max_tokens=4096,
    betas=["mcp-client-2025-11-20"],
    mcp_servers=[{
        "type": "url",
        "url": "https://api.puntersedge.online/mcp",
        "name": "puntersedge",
        "authorization_token": os.environ["PUNTERSEDGE_API_KEY"],
    }],
    tools=[{"type": "mcp_toolset", "mcp_server_name": "puntersedge"}],
    messages=[{"role": "user", "content": "Best price for every runner in the next race at Randwick?"}],
)

OpenAI Responses API

The key goes in the tool's authorization field. Every tool is read-only, so approval can be skipped.

import os

from openai import OpenAI

client = OpenAI()
response = client.responses.create(
    model="<your OpenAI model>",
    tools=[{
        "type": "mcp",
        "server_label": "puntersedge",
        "server_url": "https://api.puntersedge.online/mcp",
        "authorization": os.environ["PUNTERSEDGE_API_KEY"],
        "require_approval": "never",
    }],
    input="Which runners are firming across several bookmakers right now?",
)

Cursor

.cursor/mcp.json in the project, or ~/.cursor/mcp.json for every project:

{
  "mcpServers": {
    "puntersedge": {
      "url": "https://api.puntersedge.online/mcp",
      "headers": { "X-API-Key": "YOUR_KEY" }
    }
  }
}

VS Code and GitHub Copilot

.vscode/mcp.json. VS Code prompts for the key once and stores it, so it stays out of the file:

{
  "inputs": [
    { "type": "promptString", "id": "puntersedge-key", "description": "PuntersEdge API key", "password": true }
  ],
  "servers": {
    "puntersedge": {
      "type": "http",
      "url": "https://api.puntersedge.online/mcp",
      "headers": { "X-API-Key": "${input:puntersedge-key}" }
    }
  }
}

Claude Desktop

Claude Desktop's Connectors screen signs in with OAuth, so pass the key through the mcp-remote bridge in claude_desktop_config.json (Settings → Developer → Edit Config), then restart Claude Desktop. Needs Node.js.

{
  "mcpServers": {
    "puntersedge": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://api.puntersedge.online/mcp",
               "--header", "X-API-Key:${PUNTERSEDGE_API_KEY}"],
      "env": { "PUNTERSEDGE_API_KEY": "YOUR_KEY" }
    }
  }
}

Microsoft Copilot Studio

1. In your agent, open Tools → Add a tool → New tool → Model Context Protocol.
2. Server URL: https://api.puntersedge.online/mcp.
3. Authentication: API key, sent as a header named X-API-Key.
4. Create the connection with your key, then add the tools you want the agent to use.

n8n

1. Add an MCP Client Tool node to your AI Agent.
2. Endpoint: https://api.puntersedge.online/mcp; transport: HTTP Streamable.
3. Authentication: Header Auth with the name X-API-Key and your key as the value.

LangChain and LangGraph

With langchain-mcp-adapters, the tools load like any other LangChain tools:

import os

from langchain_mcp_adapters.client import MultiServerMCPClient

client = MultiServerMCPClient({
    "puntersedge": {
        "transport": "streamable_http",
        "url": "https://api.puntersedge.online/mcp",
        "headers": {"X-API-Key": os.environ["PUNTERSEDGE_API_KEY"]},
    }
})
tools = await client.get_tools()

Any HTTP client

Every request stands alone, so one POST is a complete tool call:

curl -s https://api.puntersedge.online/mcp \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $PUNTERSEDGE_API_KEY" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call",
       "params":{"name":"odds_board","arguments":{"num_races":3,"country":"AU"}}}'