This project addresses the challenge of understanding why certain Xiaohongshu (XHS) notes fail to gain traction. It offers a sophisticated, data-driven solution for content creators and analysts by providing quantitative diagnostics, actionable optimization strategies, simulated user comments, and AI-generated content rewrites. The primary benefit is moving beyond subjective guesswork to objective, AI-powered insights for content improvement.

How It Works

NoteRx employs a multi-agent collaborative diagnosis engine. Its core approach involves self-trained models: a quantitative prediction engine (Model A) trained on 874 real XHS notes for a 5-dimension scoring system, a baseline knowledge graph derived from notes and comments, and a comment persona engine analyzing 2465 comments. The diagnosis proceeds through four stages: data-driven baseline training, Model A's rapid initial assessment, a multi-agent debate involving 4 diagnostic agents and 1 judge agent, and finally, an AI-driven optimization loop that generates and re-scores content revisions. This multi-agent debate architecture allows for parallel diagnosis, cross-examination, and a synthesized final judgment, offering a more robust analysis than single-model approaches.

Quick Start & Requirements

To get started, clone the repository (git clone https://github.com/jiangmuran/noterx.git && cd noterx), configure backend API keys by copying .env.example to .env (cp .env.example backend/.env), and then install and launch the application using make install && make data && ./start.sh. The application can be accessed at http://localhost:5173. Prerequisites include setting up API keys for backend services.

Highlighted Details

• Achieved "Topic Star #1" in the Xiaohongshu Hackathon.
• Developed by a team of four 13-year-old middle school students in 48 hours.
• Utilizes a multi-modal input system supporting screenshots and video uploads.
• Features a 5-AI expert debate system for in-depth, cross-validated diagnosis.
• Provides a simulated comment section, including arguments and replies, and generates 3 high-scoring rewrite options.

Maintenance & Community

The project was developed by a student team (PageOne: Jiang Muran, Yang Xizhe, Chen Yuxia, Lü Sitong) for a hackathon. No explicit information regarding ongoing maintenance, community channels (like Discord or Slack), or a long-term roadmap is provided in the README.

Licensing & Compatibility

The project is licensed under the Apache License 2.0. This license is permissive and generally compatible with commercial use and linking in closed-source projects.

Limitations & Caveats

As a hackathon project developed under tight time constraints by young students, NoteRx may represent a proof-of-concept rather than a fully production-hardened system. The README does not detail ongoing support, bug-fixing schedules, or future development plans. Backend functionality relies on external API keys, indicating potential third-party service dependencies.