Mentat

Mentat - Obsidian Plugin

AI-powered personal assistant for Obsidian that helps organize notes, create knowledge links, and build your personal knowledge base.

Features

• Auto Classification: Automatically classify and tag your notes with AI
• Smart Link Suggestions: Get intelligent suggestions for linking related notes
• AI Chat: Chat with your knowledge base using RAG (Retrieval-Augmented Generation)
• Knowledge Graph: Visualize your notes in an interactive 3D graph
• Review System: Spaced repetition algorithm for effective knowledge review

AI Provider Support

Mentat supports multiple AI providers:

• OpenAI-compatible APIs: OpenAI, DeepSeek, and any OpenAI-compatible endpoint
• Anthropic: Native support for Claude models
• Ollama: Local AI models for privacy and speed

Installation

For Development

1. Clone this repository into your Obsidian vault's plugins folder:

   cd /path/to/your/vault/.obsidian/plugins
   git clone https://github.com/Secbone/obsidian-mentat mentat
   cd mentat
   

2. Install dependencies:

   npm install
   

3. Build the plugin:

   npm run build
   

4. Reload Obsidian and enable "Mentat" in Settings → Community plugins

For Users

1. Open Obsidian Settings
2. Go to Community plugins and browse
3. Search for "Mentat"
4. Install and enable the plugin

Configuration

Setting up AI Providers

1. Go to Settings → Mentat
2. Click "Add AI Provider"
3. Configure your provider:
   • Name: Give it a descriptive name
   • Type: Select OpenAI, Anthropic, or Ollama
   • API Key: Your API key (if required)
   • Base URL: Custom endpoint URL (for OpenAI-compatible APIs)
   • Model: The model to use

Task Routing

You can assign different AI providers to specific tasks:

• Embedding: For generating note embeddings (usually Ollama)
• Classification: For classifying and tagging notes
• Linking: For suggesting links between notes
• Chat: For conversational queries
• Review: For generating review questions

Optional Integrations

• OpenCode: Enable advanced automation features
• Obsidian Skills: Register and use Skills commands

Usage

Commands

• /classify-note: Classify the current note
• /suggest-links: Get link suggestions for the current note
• /open-chat: Open AI chat interface
• /open-graph: View knowledge graph
• /start-review: Begin a review session

Skills Integration

When Skills integration is enabled, you can use:

• /pa-classify: Classify note (callable by other skills)
• /pa-link: Suggest links (callable by other skills)
• /pa-chat: Chat interface (callable by other skills)

Development

Project Structure

src/
├── main.ts                 # Plugin entry point
├── settings/               # Settings management
├── ai/                     # AI providers and routing
├── features/               # Core features
│   ├── classification/
│   ├── linking/
│   ├── chat/
│   ├── graph/
│   └── review/
├── indexing/               # Vector indexing system
├── ui/                     # UI components
└── types/                  # TypeScript definitions

Building

npm run dev    # Development mode with watch
npm run build  # Production build
npm run lint   # Run ESLint

Testing

npm test

Performance

• Optimized for vaults with 100-1000 notes
• Batch processing to avoid UI blocking
• Incremental indexing for fast updates
• LRU cache for embeddings

Privacy

• Use Ollama for local, private AI processing
• All data stays in your vault
• No data sent to cloud unless you configure cloud AI providers

License

MIT

Support

• GitHub Issues: Report bugs or request features
• Documentation: Full documentation

Roadmap

• Phase 1: Project setup ✓
• Phase 2: AI providers integration
• Phase 3: Indexing system
• Phase 4: Auto classification
• Phase 5: Link suggestions
• Phase 6: Chat interface
• Phase 7: Knowledge graph
• Phase 8: Review system
• Phase 9: UI/UX polish
• Phase 10: Testing and release

Credits

Built with:

• Obsidian
• Anthropic Claude
• OpenAI
• Ollama
• 3D Force Graph