howrisky

HowRisky MCP Server provides Monte Carlo risk analysis tools for AI agents, including portfolio risk metrics (CVaR, VaR, ruin probability), startup equity modeling, real estate cash flows, and Kelly criterion betting. The server exposes a suite of tools that can be discovered by your AI agent, then invoked with the appropriate parameters to produce risk metrics, forecasts, and scenario analyses. Start by obtaining an API key from HowRisky and configuring your MCP stack with the standard config so your AI can reach the HowRisky tools via tools/list and specific tool calls. When ready, you can request calculations such as CVaR for a portfolio, simulate future timelines, compare portfolios, convert text to portfolio allocations, or model illiquid assets like startups and real estate.

Once configured, your AI can discover available tools with tools/list, then call methods like calculate_portfolio_risk, simulate_future_timelines, or add_startup with the correct inputs. The results will include risk metrics (e.g., CVaR, VaR, ruin probability), percentiles, and scenario insights. You can use the HTTP endpoint for custom integrations by posting JSON-RPC requests to the MCP server endpoint, enabling remote tools access for web apps or other services.

Prerequisites:

• Node.js (recommended: current LTS) and npm installed on your machine
• Access to the HowRisky MCP server package (via npm registry)
• An HowRisky API key

Step-by-step:

1. Install Node.js and npm if you haven’t already.

   • On macOS with Homebrew: brew install node
   • On Windows: download the installer from nodejs.org
   • On Linux: use your distribution’s package manager (e.g., apt, dnf)

2. Get your MCP config configured. Save the standard config snippet from the README into your AI tool’s MCP configuration file. Example:

{
  "mcpServers": {
    "howrisky": {
      "command": "npx",
      "args": ["-y", "howrisky-mcp-server"],
      "env": {
        "HOWRISKY_API_KEY": "your-api-key-here"
      }
    }
  }
}

3. Obtain an API key from HowRisky and replace the placeholder in the config or set HOWRISKY_API_KEY in your environment.

4. Run or initialize the MCP server via your MCP manager (rely on your tool’s procedure to load the mcpServers section). If testing locally, ensure npx is available and run the server with the provided config.

5. Test the endpoint (optional). To call the remote HTTP endpoint, see the example under the Remote Server section of the README.

Tips and considerations:

• Keep your HOWRISKY_API_KEY secure. Do not commit it to public repos.
• If your environment blocks npx downloads, you can pre-install howrisky-mcp-server locally or mirror the package in your registry.
• When using the HTTP endpoint, include the header X-API-Key with your API key for authentication.
• Review usage limits and pricing on HowRisky to understand call quotas and costs.
• If you experience latency or connection issues, check your network firewall and ensure the MCP gateway can reach https://mcp.howrisky.ai or your specified endpoint.