Setup

Connect your AI to Kalshi.

A few minutes, start to finish. Works with Claude, ChatGPT, Cursor, or any MCP-compatible client — your AI proposes, you approve every trade, and your funds stay in your own Kalshi account.

  1. 1

    Get your Kalshi API keys

    In Kalshi, open Settings → API and generate an API key. You'll get a Key ID and an RSA private key — keep the private key somewhere safe; you'll add both in the dashboard.

    Open Kalshi API settings →
  2. 2

    Connect your AI client

    Point your MCP client at the Kalshi MCP connection. Each client is a little different — Cursor uses a config file, while Claude and ChatGPT add the connection URL as a custom connector. Pick your client below.

    ~/.cursor/mcp.json
    {
  "mcpServers": {
    "kalshi": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://kalshi-mcp.com/api/mcp"]
    }
  }
}
    1. Copy the MCP configuration below into ~/.cursor/mcp.json.
    2. Reload Cursor and open the Kalshi MCP server connection.
    3. Complete the OAuth sign-in window, then add your Kalshi API keys in the dashboard.
    4. Ask Cursor to research a Kalshi event or check your account balance.
  3. 3

    Sign in

    The first time your client connects, you'll be sent through a quick OAuth sign-in. Approve access and your client is linked to your account.

  4. 4

    Add your keys in the dashboard

    Open the dashboard and paste your Kalshi Key ID and private key. They're encrypted at rest and used server-side only — they never pass through your AI client.

    Open the dashboard →
  5. 5

    Start asking

    Ask your AI to research a market, read an order book, or check your balance. It proposes trades; you approve each one before anything is placed.

Open the dashboardSee pricing