quactuary.mcp package¶

The quActuary MCP (Model Context Protocol) Server provides integration with Claude and other LLM assistants, exposing actuarial tools through a standardized interface.

Overview¶

The MCP server allows you to use quActuary’s actuarial modeling capabilities directly within Claude conversations. It provides tools for:

Running pricing simulations
Working with probability distributions
Managing portfolios and policy terms
Performing actuarial calculations

Usage¶

Starting the MCP Server¶

# Using the command-line interface
quactuary-mcp

# Or as a Python module
python -m quactuary.mcp.server

Configuration¶

The server can be configured in Claude Desktop’s settings:

{
  "mcpServers": {
    "quactuary": {
      "command": "python",
      "args": ["-m", "quactuary.mcp.server"],
      "env": {}
    }
  }
}

Submodules¶

quactuary.mcp.server module¶

quactuary.mcp.tools module¶

quactuary.mcp.resources module¶

quactuary.mcp.prompts module¶

quactuary.mcp.categories module¶

quactuary.mcp.formats module¶

quactuary.mcp.config module¶

quactuary.mcp.base module¶

Tool Categories¶

The MCP server organizes tools into the following categories:

Pricing Tools

run_pricing_simulation: Execute pricing model simulations
calculate_premium: Calculate premiums for policies
estimate_reserves: Estimate reserves for portfolios

Distribution Tools

create_distribution: Create frequency or severity distributions
fit_distribution: Fit distributions to data
sample_distribution: Generate samples from distributions

Portfolio Tools

create_portfolio: Build insurance portfolios
analyze_portfolio: Analyze portfolio metrics
optimize_portfolio: Optimize portfolio allocations

Utility Tools

calculate_metrics: Compute actuarial metrics
generate_reports: Generate actuarial reports
validate_inputs: Validate input parameters

Examples¶

Using MCP Tools in Claude¶

Once the MCP server is running and configured, you can use it in Claude conversations:

User: Run a pricing simulation for a GL policy with Poisson(3.5) frequency
      and Exponential(1000) severity, with a $1M deductible.

Claude: I'll run a pricing simulation for your GL policy using the MCP tools...
        [Uses run_pricing_simulation tool]

Creating Custom Tools¶

You can extend the MCP server with custom tools:

from quactuary.mcp.base import QuactuaryTool
from quactuary.mcp.categories import ToolCategory

class CustomActuarialTool(QuactuaryTool):
    category = ToolCategory.UTILITIES

    def execute(self, params):
        # Your custom implementation
        return {"result": "Custom calculation"}