home / mcp / novel processing mcp server

Novel Processing MCP Server

Provides an MCP server for intelligent novel segmentation and bulk rewrite task management with a Web UI and Cursor integration.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "imohuan-novel-segmentation-mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "<FASTMCP_DIR>",
        "run",
        "main.py",
        "--work-dir",
        "<WORK_DIR>"
      ]
    }
  }
}

You have a FastMCP-based toolchain designed to split long novels into precise segments, manage bulk rewrite tasks, and run as an MCP service from your local environment. It provides a Web UI for easy interaction and MCP integration to automate task retrieval and processing through a Cursor editor. This guide teaches you how to install, run, and use the local MCP server and its task workflows.

How to use

You will run the local MCP server and connect to it from your MCP client to manage and process novel rewriting tasks. The workflow includes loading a novel, splitting it by token counts, creating batch rewrite tasks, and executing those tasks through the MCP interface to produce rewritten outputs.

How to install

Prerequisites you need before installation:

# Install Python if not present
# Then install and run the local MCP server as described below

Two ways to deploy the MCP server locally

You can run the MCP service using UV to manage dependencies and execution, or run directly with Python. Choose one of the configurations below and place them in your mcp.json file.

{
  "mcpServers": {
    "novel-processor-local": {
      "command": "uv",
      "args": [
        "--directory",
        "<FASTMCP_DIR>",
        "run",
        "main.py",
        "--work-dir",
        "<WORK_DIR>"
      ]
    }
  }
}

{
  "mcpServers": {
    "novel-processor-local": {
      "command": "python",
      "args": ["<FASTMCP_DIR>\\main.py", "--work-dir", "<WORK_DIR>"]
    }
  }
}

Starting the Web interface locally

Start the Web UI to interact with the MCP server and manage projects, tasks, and logs.

# If using Python to run the web interface
python app.py --work-dir "<WORK_DIR>"

# Or, if you prefer UV-based workflow
uv run app.py --work-dir "<WORK_DIR>"

Available tools

split_novel

Intelligent segmentation of long texts by token counts across multiple encodings to produce accurate, standalone segments.

batch_rewrite

Create and manage bulk rewrite tasks with automatic detection and retry mechanisms, mapped to prompts and targets.

prompt_manager

Manage and configure prompts centrally in .batch_task/prompts with supported formats like .md and .txt.

web_interface

Provide a visually friendly Web UI to interact with the MCP server, projects, and tasks.

mcp_integration

Seamless integration with Cursor editor to fetch tasks, rewrite content according to prompts, and write back results.