rednote-analyzer

RedNote Analyzer MCP is a Python-based MCP server that enables AI assistants to search, analyze, and generate content for Xiaohongshu (Little Red Book). It exposes a set of RedNote tools (e.g., rednote_search_notes, rednote_get_note_detail, rednote_analyze_note, rednote_extract_patterns, rednote_generate_post, rednote_rewrite_in_style) that your AI agent can call to perform content research, pattern analysis, and post generation within the RedNote ecosystem. After configuring the MCP (via uvx in this case) in your project or tool integration, you simply direct your AI to use the available commands and let the MCP proxy handle the underlying web requests, rate limits, and data extraction.

Prerequisites:\n- Python 3.11+\n- uv (or pipx) to run MCP servers via the MCP protocol, as demonstrated here with uvx.\n- A working installation of the RedNote Analyzer MCP package (the server itself is provided by the project).\n\nStep-by-step:\n1) Install the MCP server package (via uvx-ready flow): ensure you have uvx installed and accessible in your environment.\n2) Install the RedNote Analyzer MCP component (the project/package name referenced in the README, e.g., rednote-analyzer-mcp). For example, using uvx you would install or run the server binary/module as described in the project docs.\n3) Log in to Xiaohongshu as required by the tool: run rednote-login in the terminal to authenticate and save cookies.\n4) Create an MCP configuration file in the root of your project (or use your preferred integration method) defining the server entry, for example: the content shown below.\n5) Start using the MCP via your AI tooling. Ensure that your AI tool imports the mcp server and points to the uvx-backed server with the appropriate environment variables (REDNOTE_ADAPTER, REDNOTE_HEADLESS).

Tips and common considerations:\n- The MCP relies on Xiaohongshu authentication; ensure you log in with rednote-login before making queries.\n- The tool enforces a rate limit (10 requests per minute) with a minimum 3-second delay between requests; respect these limits to avoid blocks.\n- If you encounter issues, check that REDNOTE_ADAPTER is set to playwright (or your preferred adapter) and that REDNOTE_HEADLESS is set to true if you want headless operation.\n- The environment can be adjusted per integration (CLI, editor extensions, or code assistants) using the same mcpServer configuration pattern.\n- If you switch to a different rendering or browser automation layer, update REDNOTE_ADAPTER accordingly.\n- To upgrade the MCP server, use your environment’s upgrade path (e.g., pipx upgrade rednote-analyzer-mcp).