## QAnon is a dangerous cult. This archive is for research purposes only, and I do _not_ endorse any material in this repo. # Q-Anon Posts/Drops MCP Server ![](https://badge.mcpx.dev?type=server 'qanon-mcp') [![smithery badge](https://smithery.ai/badge/@jkingsman/qanon-mcp-server)](https://smithery.ai/server/@jkingsman/qanon-mcp-server) An MCP (Model Context Protocol) server that provides access to a dataset of Q-Anon posts for anthropological/sociological research. This server allows AI assistants like Claude to search, filter, and analyze the Q-Anon drops. Posts are drawn from https://github.com/jkingsman/JSON-QAnon. You can learn more about how the source data was composed there, as well as find alternate formats, schemas, etc. ### Warning: This tool was entirely vibe coded. Use at your own risk. ## Prerequisites - Python 3.10 or higher - `uv` package manager - Claude Desktop (for Claude integration) ## Installation This tool is compatible with `uvx` and doesn't need to be cloned/installed. ### Installing via Smithery To install qanon-mcp-server for Claude Desktop automatically via [Smithery](https://smithery.ai/server/@jkingsman/qanon-mcp-server): ```bash npx -y @smithery/cli install @jkingsman/qanon-mcp-server --client claude ``` ### Manual 1. Clone or download this repository to your local machine 2. Install the required packages using `uv`: ```bash uv pip install -e . ``` ## Usage You can run the server directly with `uvx`: ```bash uvx qanon_mcp ``` ## Claude Desktop Integration To use this MCP server with Claude Desktop: 1. Make sure you have [Claude Desktop](https://claude.ai/download) installed 2. Open the Claude menu and select "Settings..." 3. Click on "Developer" in the left-hand bar and then "Edit Config" 4. Add the following configuration to the `claude_desktop_config.json` file: ```json { "mcpServers": { "qanon_mcp": { "command": "uvx", "args": [ "qanon_mcp" ] } } } ``` or, if you don't have `uvx` installed: ```json { "mcpServers": { "qanon_mcp": { "command": "uv", "args": [ "tool", "run", "qanon_mcp" ] } } } ``` 5. Save the file and restart Claude Desktop 6. Start a new conversation in Claude Desktop 7. You should see a hammer icon in the input box, indicating that tools are available ## Features ### Resources - `qanon://posts/count` - Get the total number of posts - `qanon://posts/{post_id}` - Access a specific post by ID - `qanon://posts/raw/{post_id}` - Get the raw JSON data for a specific post - `qanon://authors` - List all unique authors - `qanon://stats` - Get dataset statistics ### Tools - **get_post_by_id_tool** - Retrieve a specific post by its ID - **search_posts** - Find posts containing specific keywords or phrases - **get_posts_by_date** - Retrieve posts from a specific date range - **get_posts_by_author_id** - Find posts by a specific author ID - **analyze_post** - Get detailed analysis of a specific post including references and context - **get_timeline_summary** - Generate a chronological timeline, optionally within a date range - **word_cloud_by_post_ids** - Generate a word frequency analysis for posts within a specified ID range - **word_cloud_by_date_range** - Generate a word frequency analysis for posts within a specified date range ## Example Queries for Claude Once the MCP server is connected to Claude Desktop, you can ask questions like: - "How many Q-Anon posts are in the dataset?" - "Search for posts that mention 'storm'" - "Show me posts from October 2020" - "Analyze post #3725" - "Create a timeline of Q-Anon posts from 2018" - "Generate a word cloud for Q-Anon posts between January and March 2019" - "Get the raw data for post #4500" - "What are the most common words used in posts #1000-2000?" ## Troubleshooting - If Claude Desktop doesn't show the hammer icon, check your configuration and restart Claude Desktop - Ensure the `posts.json` file is in the same directory as the script - Check the output in the terminal for any error messages - Make sure you're using the absolute path to the script in your Claude Desktop configuration