home / mcp / semantic frame mcp server

Semantic Frame MCP Server

Provides token-efficient semantic analysis of numerical data via MCP, exposing the describe_data tool for MCP clients.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "anarkitty1-semantic-frame": {
      "command": "mcp",
      "args": [
        "run",
        "semantic_frame.integrations.mcp:mcp"
      ]
    }
  }
}

Semantic Frame MCP Server enables you to run the Semantic Frame analysis engine locally and expose its capabilities to MCP-compatible clients. You can connect tools like ElizaOS or Claude Desktop to analyze data series and DataFrames through a lightweight, standardized interface, enabling fast, deterministic insights without exposing raw data to external services.

How to use

You run the MCP server locally to provide data analysis capabilities to MCP clients. Start the server using the command shown below, then connect your MCP client to the exposed tool. The client can request a single series analysis or batch analyses for multiple columns, receiving structured results that you can incorporate into your workflow.

How to install

Prerequisites: you need Python and a package manager to install the MCP package. Ensure you have Python 3.8+ and pip available on your system.

pip install semantic-frame[mcp]

Notes

This server exposes a single MCP tool named describe_data that MCP clients can use to analyze data. The server runs as a local process and does not require a remote URL. Ensure you run the exact command shown to start the MCP server so clients can discover and use describe_data.

Available tools

describe_data

Analyze a single data series via MCP with deterministic narrative output.