Subnet Calculator MCP Server

subnet-calculator-mcp is a production-ready Model Context Protocol server that delivers reliable IPv4 subnet planning utilities to LLM-powered assistants. It eliminates tedious manual math by exposing fast, well-tested tools for subnet sizing, wildcard mask generation, gateway selection, and host validation.

Why Use This Server?

• 🔢 Convert host requirements into accurate subnet masks and CIDR prefixes

• 🌐 Produce Cisco-friendly OSPF wildcard masks and network statements

• ✅ Verify whether an IP belongs to a subnet, including gateway hints and address position

• 🔄 Reverse-calculations from dotted masks or find the Nth usable address instantly

• ⚙️ Built on the official Python MCP SDK with thorough type hints, validation, and tests

Related MCP server: NetworkCalc MCP Server

Install & Run

The package is published on PyPI: https://pypi.org/project/subnet-calculator-mcp/

Install with pip

pip install subnet-calculator-mcp
python -m subnet_calculator_mcp.server

Run instantly (no install)

uvx subnet-calculator-mcp

Install as a reusable CLI

uv tool install subnet-calculator-mcp
# later
subnet-calculator-mcp

Client Configuration

Claude Desktop / Claude for Windows

Add the server to claude_desktop_config.json.

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "subnet-calculator": {
      "command": "uvx",
      "args": ["subnet-calculator-mcp"]
    }
  }
}

If you installed via uv tool install or pip, change the command to "subnet-calculator-mcp" and omit args.

Generic mcp.json

Many editors (Cursor, Windsurf, etc.) use an mcp.json file. Add the following entry:

{
  "subnet-calculator": {
    "command": "uvx",
    "args": ["subnet-calculator-mcp"],
    "env": {}
  }
}

Set command to "subnet-calculator-mcp" if the binary is installed globally.

Tool Reference

calculate_subnet

• Inputs: network_base (IPv4 address), hosts_needed (int), optional return_format ("detailed" or "simple").

• Returns: CIDR prefix, dotted mask, wildcard mask, usable range, broadcast, binary representations, total/usable host counts.

• Use it for: deriving the smallest subnet that can host the requested number of usable addresses.

calculate_wildcard_mask

• Inputs: ip_address (any IP in the subnet), cidr_prefix (0–32), optional include_ospf_command (bool).

• Returns: network address, mask, wildcard mask, binary wildcard string, and optional network ... area 0 statement.

• Use it for: creating OSPF configurations or ACLs that rely on wildcard masks.

validate_ip_in_subnet

• Inputs: ip_address, network (CIDR string), optional return_gateway (bool).

• Returns: membership flag, mask, prefix, network/broadcast flags, usability, likely gateway, position index, and remaining usable addresses.

• Use it for: quickly verifying assignments and identifying first-hop router addresses.

calculate_subnet_from_mask

• Inputs: ip_address, subnet_mask (dotted decimal).

• Returns: network boundary, prefix, wildcard mask, usable range, broadcast, and host counts.

• Use it for: analysing legacy configurations that provide dotted masks instead of CIDR notation.

get_nth_usable_ip

• Inputs: network (CIDR string), position (1-based index).

• Returns: IP address at that position, whether it is last usable, total usable hosts, and associated network address.

• Use it for: allocating deterministic host positions (first server, second router, etc.).

Example invocation:

{
  "tool": "calculate_subnet",
  "arguments": {
    "network_base": "172.16.0.16",
    "hosts_needed": 14
  }
}

Development Workflow

git clone https://github.com/sanjayshreeyans/subnet-calculator-mcp.git
cd subnet-calculator-mcp

# Install dependencies (dev group includes pytest, mypy, ruff, black)
uv sync --group dev

# Run the automated test suite
uv run pytest

# Type-check, lint, and format
uv run mypy src
uv run ruff check src tests
uv run black --check src tests

# Build distribution artifacts
uv build

License

MIT License – see LICENSE for details.