Tokenmizer

Overview

Tokenmizer is a Model Context Protocol server that provides graph-backed session memory for language models. It enables you to checkpoint and resume conversations, analyze files, and track token-savings statistics through MCP-connected tools that integrate directly into your AI assistant workflow.

The server runs locally as a Python package installed via uvx tokenmizer, communicating over stdio with your MCP client. It requires an API key for authentication. The latest release is v0.3.1, distributed under the MIT license with source code available at Shweta-Mishra-ai/tokenmizer on GitHub.

Tokenmizer is listed in the Memory category on mcp.site and works with Claude Desktop, Claude Code, Cursor, VS Code, and other MCP clients. If you maintain this project, you can claim this listing to verify ownership and keep information up to date.

Installation

// claude_desktop_config.json
{
  "mcpServers": {
    "tokenmizer": {
      "command": "uvx",
      "args": [
        "tokenmizer"
      ]
    }
  }
}