mcp-document-system

# MCP Document System A Python-based document management system built on the **Model Context Protocol (MCP)**. Interact with documents through structured tools, resources, and prompts — with LLM-powered query support via OpenRouter. --- ## Features - Document resource system (`list_docs`, `fetch_doc`) - Tool-based document operations — read and edit document contents - LLM integration via OpenRouter for natural language queries - Fully functional MCP server with Inspector support - CLI-based interactive client --- ## Architecture ``` CLI (main.py) ↓ MCP Client (mcp_client.py) ↓ MCP Server (mcp_server.py) ↓ Tools + Resources (tools.py, core/) ↓ Documents ``` --- ## Screenshots ### MCP Inspector — Tools  ### MCP Inspector — Resources  ### MCP Inspector — Prompts  ### CLI Interaction  --- ## Setup ### 1. Clone the repository ```bash git clone https://github.com/<your-username>/mcp-document-system.git cd mcp-document-system ``` ### 2. Create a virtual environment ```bash python -m venv .venv .venv\Scripts\activate # Windows # source .venv/bin/activate # macOS/Linux ``` ### 3. Install dependencies ```bash pip install mcp anthropic python-dotenv ``` ### 4. Configure environment variables Create a `.env` file in the project root: ```env OPENROUTER_API_KEY=your_api_key_here OPENROUTER_MODEL=qwen/qwen-2.5-72b-instruct ``` --- ## Usage ### Run the MCP Server (for Inspector) ```bash mcp dev cli_project/mcp_server.py ``` Then open the Inspector at: `http://localhost:6274` ### Run the CLI Client ```bash python -m cli_project.main ``` **Example query:** ``` > what is the contents of report.pdf document The report details the state of a 20m condenser tower. ``` --- ## Available Tools | Tool | Input | Output | |------|-------|--------| | `read_doc_contents` | `doc_id` | Document text | | `edit_document` | `doc_id`, `old_str`, `new_str` | Updated document | --- ## Project Structure ``` cli_project/ ├── core/ │ ├── chat.py │ ├── cli.py │ ├── cli_chat.py │ ├── tools.py │ └── claude.py ├── mcp_client.py ├── mcp_server.py ├── main.py ├── pyproject.toml └── README.md screenshots/ ├── tools.png ├── resources.png ├── prompts.png └── cli.png ``` --- ## MCP Inspector Capabilities | Action | Description | |--------|-------------| | `resources/list` | List all available documents | | `resources/read` | Read a specific document | | `tools/call` | Execute a tool (read or edit) | | `prompts/list` | Inspect registered prompts | --- ## Roadmap - [ ] Database storage backend - [ ] Improved PDF/DOCX parsing - [ ] Authentication layer - [ ] Web UI --- ## Author **Priya Singh** — B.Tech CSE | AI/ML Enthusiast --- *If this project helped you, consider giving it a ⭐ on GitHub!*