SlideGen_MCP

# Automated MCP SlideGen Tool This project consists of three MCP (Model Context Protocol) servers integrated with the Claude Desktop app to power context-aware GenAI workflows. Each server specializes in a unique task: * `documentation`: Scrapes and summarizes webpages. * `ppt_generator`: Generates McKinsey-style PowerPoint slides from documents. * `pdf_analyzer`: Extracts and processes information from PDFs. The MCP servers are configured via Claude’s JSON setup and managed with [`uv`](https://github.com/astral-sh/uv) for fast, reproducible Python environments. --- ## Quickstart ### 1. Install `uv` ```powershell powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex" ``` --- ### 2. Install and Set Up MCP Servers Clone or organize your project directory like this: ``` D:\2025\pyth\ ├── documentation\ │ └── main.py ├── ppt_generator\ │ └── main.py ├── pdf_analyzer\ │ └── main.py ``` ### 3. Install Dependencies Using `uv` Run the following inside each respective folder: #### `documentation`: ```bash uv add bs4 httpx "mcp[cli]" python-dotenv ``` #### `ppt_generator`: ```bash uv add fitz httpx "mcp[cli]" numpy pillow pymupdf python-dotenv python-pptx scikit-learn ``` #### `pdf_analyzer`: ```bash uv add httpx "mcp[cli]" nltk numpy pymupdf pypdf2 scikit-learn ``` --- ### 4. Configure Claude Desktop Update your `claude.config.json` (or the relevant config file) to point to each MCP server: ```json { "mcpServers": { "documentation": { "command": "C:\\Users\\Rosha\\.local\\bin\\uv.exe", "args": ["--directory", "D:\\2025\\pyth\\documentation", "run", "main.py"], "env": {} }, "ppt_generator": { "command": "C:\\Users\\Rosha\\.local\\bin\\uv.exe", "args": ["--directory", "D:\\2025\\pyth\\ppt_generator", "run", "main.py"], "env": {} }, "pdf_analyzer": { "command": "C:\\Users\\Rosha\\.local\\bin\\uv.exe", "args": ["--directory", "D:\\2025\\pyth\\pdf_analyzer", "run", "main.py"], "env": {} } } } ``` --- ### 5. Run the Servers In each project directory: ```bash uv run main.py ``` #### Development Mode (with Inspector) ```bash npx @modelcontextprotocol/inspector uv run main.py ``` --- ## MCP Client & Frontend (To Do) * [ ] Implement MCP Client * [ ] Create Streamlit UI Frontend --- ## Verification Once configured and running, you should see your tools listed under the **Tools** section in the Claude Desktop app. Each tool will be callable from within Claude using its context-aware interface. --- ## Project Status - You can find the video of working prototype [here](https://drive.google.com/file/d/1gxlOti9rv2PiAAQVvrZJT9qw81RM848e/view?usp=sharing). - PPT generated during test can be found [here](kuhn_presentation.html). | Component | Status | | --------------- | ----------- | | `documentation` | ✅ Completed | | `ppt_generator` | ✅ Completed | | `pdf_analyzer` | ✅ Completed | | `mcp-client` | ⏳ To Do | | Streamlit UI | ⏳ To Do | --- ## Credits Built with ❤️ by Roshan Kumar. Powered by [Model Context Protocol](https://github.com/modelcontextprotocol).