Model Context Protocol MCP Server

Comics Now! includes a native Model Context Protocol (MCP) server (mcp/comics_mcp/server.py), allowing LLM assistants (Claude Desktop, Claude Code, Cursor, Gemini CLI) to interact directly with your comic library.

MCP Overview

The MCP server communicates locally with Comics Now! as an admin, giving your AI agent the ability to:

• Curate reading orders and build reading lists automatically based on comic book crossover events.
• Visually identify un-tagged comics from native cover page scans.
• Run read-only SQLite queries across your library.
• Safely call REST endpoints with built-in protection against destructive file operations (COMICS_ALLOW_DESTRUCTIVE=0 by default).

MCP Setup & Launcher

1. Install Python dependencies:

   cd mcp
   python3 -m venv .venv
   ./.venv/bin/pip install -r requirements.txt
   cp .env.example .env

2. Start the server (default stdio mode):

   bash mcp/run.sh
   # Or via npm
   npm run mcp

Network mode: The default stdio transport has no listening port. For network transports (e.g. remote use), set MCP_TRANSPORT=sse (or streamable-http), plus MCP_HOST (default 127.0.0.1) and MCP_PORT (default 8000) in mcp/.env — run.sh loads that file automatically.

Client Configuration

Claude Code:

claude mcp add comics-now -- /path/to/comics-now/mcp/run.sh

Claude Desktop (claude_desktop_config.json):

{
  "mcpServers": {
    "comics-now": {
      "command": "/path/to/comics-now/mcp/run.sh"
    }
  }
}

Gemini CLI (~/.gemini/settings.json):

{
  "mcpServers": {
    "comics-now": {
      "command": "/path/to/comics-now/mcp/run.sh"
    }
  }
}

AI Capabilities

• Visual Cover Tagging (get_cover_image, tag_workflow, set_tags): The AI extracts native resolution cover art, inspects logos, issue numbers, and trade dress, searches the web to verify the exact issue, and writes ComicInfo metadata.
• Reading List Generation (create_reading_lists_bulk): Ask the AI: "Create a reading list for the Batman: Knightfall crossover in chronological order." The AI queries your library, resolves comic IDs, and builds the sorted reading list automatically.
• Safe SQL Queries (query_db): Runs read-only SELECT / PRAGMA queries to inspect library statistics, missing issues, or unread counts.