This project provides a tool to automatically generate .mdc (Markdown Cursor) rule files for libraries, aimed at enhancing the Cursor IDE experience. It leverages semantic web search via Exa and LLM-based content generation (Gemini, OpenAI, or Anthropic) to create detailed, structured documentation for library usage best practices.

How It Works

The generator script processes a rules.json file containing library metadata. For each library, it uses Exa to perform semantic searches for best practices and relevant information on the web. This gathered data is then fed to an LLM to generate comprehensive .mdc files. The process supports parallel execution, progress tracking, and smart retries for failed libraries, optimizing the generation workflow.

Quick Start & Requirements

• Install: Clone the repository and run uv sync to install dependencies.
• Prerequisites: Python 3.8+, uv for dependency management, and API keys for Exa and an LLM provider (Gemini, OpenAI, or Anthropic).
• Setup: Create a .env file with your API keys.
• Usage: Run uv run src/generate_mdc_files.py.
• Docs: Project Repository

Highlighted Details

• Generates .mdc files for Cursor IDE from structured JSON input.
• Utilizes Exa for semantic web search to gather library best practices.
• Employs LLMs (Gemini, OpenAI, Anthropic) for content creation.
• Supports parallel processing, progress tracking, and smart retries.

Maintenance & Community

This is a community-driven initiative. Contributions are welcome via pull requests.

Licensing & Compatibility

• License: MIT License.
• Compatibility: Permissive license suitable for commercial use and integration with closed-source projects.

Limitations & Caveats

The CLI tool for finding and installing rules is marked as deprecated. The project relies on external API keys, which may incur costs and require careful management.