AI-powered tagging system for organizing AI conversations and technical knowledge in Obsidian.
Magus Mark provides intelligent tagging and organization for AI chat exports and technical knowledge, with a focus on making your knowledge base more searchable and navigable. The system implements a carefully designed taxonomy with fixed categories to ensure tag consistency and utility.
- AI-Powered Tagging: Automatically analyze and tag conversations with consistent, meaningful tags
- Cross-Platform Integration: Works in Obsidian, VS Code, Cursor, and via CLI
- Hierarchical Tag System: Sophisticated classification system with hierarchical taxonomies
- Cost Management: Control API usage with budget limits and optimization
- Customizable: Configure tagging behavior, models, and taxonomy
Magus Mark is a suite of tools that work together:
The shared core library that powers all components:
- Sophisticated classification system with hierarchical taxonomies
- Precision-engineered prompts for OpenAI integration
- Optimized token usage for cost efficiency
Seamless integration with Obsidian:
- Tag documents directly within Obsidian
- Manage tags through a dedicated interface
- Visualize knowledge connections
- Learn more about the Obsidian Plugin
Integration with VS Code and Cursor:
- Connect VS Code with your Obsidian vaults
- Special integration with Cursor AI
- Tag code snippets and conversations
- Knowledge retrieval based on coding context
- Learn more about the VS Code Extension
Process files and directories from the command line:
- Batch processing capabilities
- Interactive and scriptable modes
- Advanced cost management
- Performance benchmarking
- Learn more about the CLI
- Node.js 18 or higher
- pnpm (for workspace management)
- OpenAI API key
# Clone the repository
git clone https://github.com/magus-mark/magus-mark.git
cd magus-mark
# Install dependencies
pnpm install
# Build all packages
pnpm build- Obsidian Plugin: Install from Obsidian Community Plugins or manually
- VS Code Extension: Install from VS Code Marketplace or manually
- CLI Tool: Install via npm or use from source
See individual component READMEs for detailed installation instructions.
This project uses a modern TypeScript workspace with pnpm, ESLint flat config, and project references:
# Build all components
pnpm build
# Run tests
pnpm test
# Lint
pnpm lint
# Build specific component
pnpm --filter @magus-mark/core buildThe project uses GitHub Actions for CI/CD:
- Component-specific CI workflows in
apps/<component>/.github/workflows/ - Repository-level workflows in
.github/workflows/ - Automated release process for all components
- Accessibility testing for UI components
For detailed documentation, see:
Contributions are welcome! Please read the contributing guidelines first.
This project is licensed under the MIT License - see the LICENSE file for details.
- Built on the Obsidian Plugin API and VS Code Extension API
- Powered by OpenAI's language models
- Inspired by the needs of knowledge workers and developers