MCP Server (MCP Protocol Compliant)

Overview

This server is now a standards-compliant Model Context Protocol (MCP) server. It exposes document generation as an MCP tool, allowing integration with AI hosts (e.g., Copilot, Cursor, Claude Desktop) via the MCP protocol.

MCP Features

• MCP Tool: generateDocument — Generate PDF, DOCX, PPTX, or XLSX files from a prompt.
• Transport: Stdio (local, default). SSE (network) can be added if needed.
• Schema Validation: Uses zod for input validation.
• Extensible: Add more tools/resources/prompts as needed.

Features

• Document Generation: Create PDF, Word, PowerPoint, and Excel files from AI-driven prompts.
• Image Embedding & Resizing: Insert and resize images in DOCX and PPTX files.
• Asynchronous Processing: Queue and process document generation jobs using BullMQ and Redis.
• Validation: Validate document layout by capturing and analyzing screenshots.
• API-First: RESTful endpoints for document generation, job status, and result retrieval.
• Extensible: Modular architecture for easy addition of new file types or tools.

Project Structure

src/
  app.js
  server.js
  config/
  controllers/
  middleware/
  public/
  routes/
  services/
  utils/
  workers/
temp_image/
validation-screenshots/
test-output/

Getting Started

Prerequisites

• Node.js (v14+ recommended)
• Redis server

Installation

1. Clone the repository:

   git clone <repository-url>
   cd mcp-server
   

2. Install dependencies:

   npm install
   

3. Set up environment variables:
   • Copy .env.example to .env and fill in the required values.

Running the Server

npm start

Running Tests

npx jest src/services/fileCreation.test.js --runInBand

API Endpoints

• POST /documents/generate — Request document generation (PDF, DOCX, PPTX, XLSX)
• GET /documents/generate/status/:jobId — Check job status
• GET /documents/generate/result/:jobId — Download generated file

File Generation & Validation

• Supports AI-driven content and layout via Gemini/OpenAI function calling.
• Image insertion with resizing is available for DOCX and PPTX.
• Layout validation uses screenshots (mocked or real, see aiFileValidationService).

Example: Image Embedding & Resizing

• Place images in temp_image/ (e.g., temp_image/blackboard-with-camera-leaf.jpg).
• Use function calls like:

  { name: 'addImage', arguments: { path: 'temp_image/blackboard-with-camera-leaf.jpg', options: { width: 150, height: 150 } } }
  

Usage

Build and Start

npm run build
npm start

MCP Host Configuration Example (stdio)

Add to your host's mcp.json (e.g., for Cursor, VS Code, Claude Desktop):

{
  "mcpServers": {
    "document-server": {
      "command": "/usr/local/bin/node",
      "args": ["/absolute/path/to/build/index.js"]
    }
  }
}

Debugging with MCP Inspector

You can test this server using the official MCP Inspector:

npx @modelcontextprotocol/inspector node build/index.js

Extending

• Add more tools/resources in src/index.ts using the MCP SDK.
• For SSE/network transport, see the MCP SDK docs and the guide.

Security & Best Practices

• Always validate and sanitize tool/resource inputs.
• Never expose sensitive data without user consent.
• For remote/SSE deployments, implement authentication and TLS.

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

License

MIT

---

This project now follows the official MCP server standards for Node.js/Typescript.