HowRisky MCP Server

Monte Carlo risk analysis for AI agents. Institutional-grade financial modeling with fat-tail distributions and proprietary KDE algorithms.

8 Tools: Portfolio risk (CVaR, ruin probability), startup equity, real estate, Kelly criterion betting, and more.

Compatible with: Claude Desktop, ChatGPT Desktop, Cursor, Windsurf, Cline, GitHub Copilot, VS Code, Codex

---

Standard Config

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

Get your free API key at: https://howrisky.ai/app/settings (100 calls/month free)

---

Getting Started

Step 1: Get your API key from https://howrisky.ai/app/settings

Step 2: Add the standard config above to your AI tool's MCP configuration

That's it! Your AI can now access Monte Carlo risk simulations.

---

Installation

<details> <summary>Claude Desktop</summary>

Edit config file:

• MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

Add the standard config above.

Restart Claude Desktop.

Test it: Ask Claude "Using HowRisky, what's the risk of a 60/40 portfolio?"

</details> <details> <summary>ChatGPT Desktop</summary>

1. Open ChatGPT Desktop Settings
2. Go to Apps & Connectors → Advanced settings
3. Enable Developer mode
4. Add MCP server configuration (use standard config above)

Restart ChatGPT Desktop.

Test it: Ask ChatGPT "Use HowRisky to calculate CVaR for 100% SPY portfolio"

</details> <details> <summary>Cursor</summary>

Add to Cursor's MCP configuration file:

Use the standard config above.

Cursor supports MCP via VS Code extension compatibility.

</details> <details> <summary>Windsurf</summary>

Add to Windsurf MCP settings:

Use the standard config above.

Windsurf's MCP integration works similarly to Cursor.

</details> <details> <summary>Cline (VS Code)</summary>

Via Cline MCP Marketplace:

1. Open Cline in VS Code
2. Search for "howrisky" in MCP Marketplace
3. Click Install
4. Enter API key when prompted

Manual Setup:

Add to VS Code Settings → Extensions → Cline → MCP Servers:

Use the standard config above.

</details> <details> <summary>GitHub Copilot / VS Code</summary>

Add to VS Code settings.json:

Use the standard config above in the MCP servers configuration section.

</details> <details> <summary>Remote Server (HTTP)</summary>

For custom integrations or web-based AI tools:

Endpoint: https://mcp.howrisky.ai

Authentication: Include X-API-Key header with your API key

Documentation: https://howrisky.ai/mcp/docs

Example:

curl -X POST https://mcp.howrisky.ai \
  -H "X-API-Key: your-api-key" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","method":"tools/list","id":1}'

</details>

---

Available Tools

Full documentation: https://howrisky.ai/mcp/docs

---

Example Usage

Once configured, ask your AI:

"Using HowRisky, calculate the risk of investing $100K in a 60/40 portfolio over 20 years"

The AI will:

1. Discover HowRisky tools via tools/list
2. Call calculate_portfolio_risk with correct parameters
3. Return CVaR, survival probability, ruin risk, and other metrics

---

Features

Fat-Tail Modeling - Gaussian models underestimate crash risk by 3-10x. Our proprietary KDE captures reality.

Comprehensive Metrics - 12 risk metrics including CVaR 95/99, VaR, ruin probability, percentiles

Private Assets - Model startups, real estate, PE funds, and high-risk gambles

Tax-Aware - 15+ countries supported (US, GB, DE, FR, IT, ES, JP, AU, CA, etc.)

Custom Scenarios - Override historical data with your own market assumptions

---

Pricing

View pricing: https://howrisky.ai/mcp/pricing

---

Support

• Issues: https://github.com/howrisky/howrisky-mcp-server/issues
• Docs: https://howrisky.ai/mcp/docs
• Email: [email protected]

---

License

Proprietary - Copyright © 2025 Diogo Seca / HowRisky.ai

You may use this software to access HowRisky MCP API. Modification and redistribution prohibited. See LICENSE for details.