This project provides a visual, AI-powered document polishing tool for writers and content creators. It allows users to break down documents into draggable cards, manage prompts as reusable cards, and orchestrate AI-driven text modifications through a visual interface, simplifying complex editing workflows.

How It Works

The tool parses Markdown documents into individual text cards. Users can create and manage prompt cards, which can be linked to text cards using a "plug-and-socket" metaphor. The system supports various AI models, including Tongyi Qianwen, DeepSeek, and Ollama, allowing for flexible AI integration. Output is generated by concatenating the processed text cards back into a Markdown format.

Quick Start & Requirements

• Install: Clone the repository and run chmod +x start.sh && ./start.sh (Linux/macOS) or sh start.sh (Windows via Git Bash/PowerShell/CMD).
• Prerequisites: Node.js, Ollama (optional for local models).
• Configuration: Copy config.example.js to config.js and set API keys for online models.
• Access: http://localhost:3000 (full mode) or http://127.0.0.1:3000 (local Ollama mode).
• Docker: Available via docker build -t prose-polish:latest . and docker run -d -p 3333:3000 -e MODE=1 --name prose-polish prose-polish:latest.
• Demo: A sample document "端午的鸭蛋" is provided for testing.

Highlighted Details

• Visual, card-based interface for document editing and AI prompt management.
• Supports multiple AI models including local Ollama integration.
• JSON import/export for prompt cards.
• "Plug-and-socket" system for connecting prompts to text blocks.

Maintenance & Community

The project encourages UI modifications through derivative projects, with instructions provided for submission. The core project will not merge UI-related pull requests to maintain stability.

Licensing & Compatibility

MIT License. Permissive for commercial use, modification, and distribution, provided license and copyright notices are retained.

Limitations & Caveats

The project explicitly states it will not merge UI-related pull requests to prevent interface clutter. Some recent updates have introduced minor behavioral changes to card interactions.