chuk -math

The Chuk MCP Math Server provides a large collection of 642 mathematical functions accessible via either STDIO or HTTP transport. It is built on the chuk-mcp-server framework and exposes a rich set of domains including arithmetic/number theory, trigonometry, linear algebra, statistics, calculus, geometry, probability, and more. You can use the server via the provided lightweight clients (uvx for a no-install experience or uv for development) and then issue RPC-style calls to list available tools or invoke specific functions. The server supports function filtering, configurable timeouts, and flexible transport options to suit local development or production deployments. For Claude Desktop integration, you can point clients to the public server or run locally with uvx to access the full 642-function capability set.

Prerequisites:

• Python 3.11 or newer
• pip (Python package installer)
• Optional: uv or uvx for transport options

Installation options:

Option A: Use uvx (no installation required; run directly from the network)

uvx chuk-mcp-math-server

Option B: Use uv (development or local install)

# Install the package from PyPI
uv pip install chuk-mcp-math-server

# Or clone and install from source
git clone https://github.com/chuk-mcp/chuk-mcp-math-server.git
cd chuk-mcp-math-server
uv sync

Option C: Traditional Python installation with pip

pip install chuk-mcp-math-server

After installation, start the server using the desired transport (STDIO or HTTP) as described in the usage section.

Notes and tips:

• Transport options: STDIO (for Claude Desktop) or HTTP (web API). The public server can be used for quick testing without local installation.
• Function filtering: Use --domains, --functions, and --exclude-functions to tailor the exposed API surface.
• Configuration: Server supports CLI flags and config files (YAML/JSON) as described in the README. You can use environment variables for container-friendly configuration.
• Claude Desktop integration: You can point Claude to the public server or run locally with uvx/uv and reference the local server, using absolute paths when needed.
• If you encounter issues with 97 prime checks or specific functions, ensure you are querying the correct function name and providing proper argument structures (e.g., {"n": 97} for is_prime).
• If using the HTTP API, you can first request tools/list to discover available functions, then call tools/call with the function name and arguments.