Unit Converter AI MCP

What unit-converter-ai-mcp MCP server does

The unit-converter-ai-mcp MCP server provides measurement conversion tools for AI agents using the Model Context Protocol. The repository specifically describes length and weight conversion, while its usage examples also reference temperature conversion. It is intended for automation workflows in which an MCP client can ask an agent to select and call a conversion operation.

The project is published as a Python package named unit-converter-ai-mcp and is licensed under MIT. It does not document a required external account, API token, or hosted service dependency.

How it works

The server uses the MCP protocol to expose conversion operations to a compatible client. Its documented transport is stdio: an MCP client starts the unit-converter-ai-mcp command and exchanges messages with the running process.

After configuration, an assistant can be prompted to use named operations such as convert_length, convert_weight, or convert_temperature. The supplied material does not define the complete list of accepted unit names, argument schemas, precision rules, or error responses, so those details should be checked in the linked API reference or package documentation before building strict application logic around them.

For local use, the unit-converter-ai-mcp MCP server can be installed from PyPI with pip or launched with uvx. No environment variables are listed in the repository instructions.

Setup and configuration

Install the package with:

pip install unit-converter-ai-mcp

The README also provides a direct stdio configuration using uvx:

{
  "mcpServers": {
    "unit-converter-ai-mcp": {
      "command": "uvx",
      "args": ["unit-converter-ai-mcp"]
    }
  }
}

The same configuration pattern is identified for Claude Desktop and for MCP client configuration generally. Alternatively, after installing with pip, run the unit-converter-ai-mcp command as the stdio server. The README mentions Smithery as another installation route, but the project’s own package installation and command are the relevant local setup options.

Tools and capabilities

Documented operations include:

• convert_length for length conversions.
• convert_weight for weight conversions.
• convert_temperature, shown in the repository’s example prompts.

The server is described as MCP-protocol compliant and provides a Python distribution through PyPI. The available material does not state whether it supports compound units, batch requests, custom conversion formulas, or configurable rounding.

Limitations and notes

The repository’s short description emphasizes length and weight, while the examples additionally mention temperature. Treat temperature support as documented usage guidance, but verify the current package behavior and supported units against the API reference before depending on it.

No supported clients are listed beyond the Claude Desktop configuration example, although any MCP client that can start a local stdio server may be suitable in principle. The supplied material does not provide version-specific compatibility requirements, runtime constraints, test instructions, or detailed tool schemas. It also includes broader MEOK ecosystem and governance references; those are separate offerings and are not required to run the unit-converter-ai-mcp MCP server.