Modular Intelligence Network for Integrated Operations Natively

A unified personal intelligence platform that runs natively on Linux, Windows, macOS, Mini PCs, and high-performance workstations. Built with Rust and Tauri 2 for native performance and a small footprint.

• No Docker Required - Native execution, zero containerization dependency
• Offline-First - Full functionality without internet
• Privacy-First - Zero telemetry unless explicitly enabled
• Local AI - Ollama-compatible local LLM integration
• Modular - Enable only what you need

• Multi-directory scanning with duplicate detection (SHA-256 hashing)
• Multi-directory file comparison
• Bulk delete/move duplicate files
• Optimized scan: hashes only size-candidates (10x faster than naive)
• Live scan progress with file count and cancel support
• Exclusion patterns for skipping directories
• Storage analytics and breakdown

• EPUB parsing with full chapter navigation (arrow keys)
• EPUB cover image extraction and display
• EPUB inline image rendering (base64 data URIs)
• Persistent book library (books survive app restarts)
• Book collections/playlists with color coding
• Annotations and knowledge base
• O'Reilly Learning SSO login via Chrome cookies

• CSV bank statement import with auto-categorization
• Portfolio tracking and expense analysis
• CAGR calculator and net worth breakdown
• Account management and transaction history

• Workout tracking and habit logging
• Nutrition tracking with calorie/macro breakdown
• AI health analysis (sends metrics to local Ollama LLM)
• Historical charts and trend visualization

• Video processing and metadata extraction
• YouTube automation and thumbnail generation

• Multi-platform publishing
• SEO optimization

• Command palette (Ctrl+K) with fuzzy search
• Dashboard with real data from all modules
• Dark/Light theme support
• App icon (blue M gradient)
• 623+ tests across the workspace

• Backend: Rust (Tokio async runtime), 14 workspace crates
• Frontend: Tauri 2 + SolidJS + TypeScript + Tailwind CSS
• Database: SQLite (rusqlite) with r2d2 connection pooling, 20+ tables
• Search: tantivy full-text search index
• AI: Ollama LLM integration, embeddings, RAG pipeline
• Encryption: AES-256-GCM via RustCrypto, Argon2id key derivation

• RAM: < 300MB baseline
• Storage: ~100MB application + user data
• OS: Linux (Ubuntu 20.04+), Windows 10+, macOS 11+
• Optional: Ollama for local AI features

sudo apt-get install -y libgtk-3-dev libwebkit2gtk-4.1-dev \
  libayatana-appindicator3-dev librsvg2-dev

# Clone
git clone https://github.com/yourusername/minion.git
cd minion

# Install frontend dependencies
cd ui && pnpm install && cd ..

# Run in development mode (builds UI + Rust, launches app)
cargo tauri dev

# Build for production (.deb and .AppImage on Linux)
cargo tauri build

Production artifacts are placed in src-tauri/target/release/bundle/.

minion/
├── crates/
│   ├── minion-core/       # Plugin trait, event bus, config, error types, task system
│   ├── minion-db/         # SQLite with connection pooling (r2d2) and migrations
│   ├── minion-crypto/     # AES-256-GCM encryption, Argon2 key derivation, vault
│   ├── minion-ai/         # Ollama LLM integration, embeddings, RAG pipeline
│   ├── minion-search/     # Full-text search via tantivy
│   ├── minion-plugins/    # Plugin loader/registry
│   ├── minion-files/      # File Intelligence module
│   ├── minion-media/      # Media Intelligence module
│   ├── minion-blog/       # Blog AI Engine module
│   ├── minion-finance/    # Finance Intelligence module
│   ├── minion-fitness/    # Fitness & Wellness module
│   ├── minion-reader/     # Book Reader module
│   ├── minion-app/        # CLI entry point
│   └── minion-integration-tests/
├── src-tauri/             # Tauri 2 backend (commands, state)
├── ui/                    # SolidJS + TypeScript frontend
│   └── src/
│       ├── components/    # Shared layout, command palette
│       └── pages/         # Dashboard, Files, Reader, Finance, Fitness, Settings
├── docs/                  # Architecture, API, security, roadmap docs
└── scripts/               # Release and version management

The Tauri layer (src-tauri/commands.rs) exposes #[tauri::command] IPC functions that the SolidJS frontend invokes. Shared app state lives in src-tauri/state.rs.

MINION stores its SQLite database and configuration in a platform-specific data directory managed by Tauri (typically ~/.local/share/minion/ on Linux).

1. Install Ollama and pull a model (e.g., ollama pull llama3.2)
2. Open Settings in the app
3. Set the Ollama URL (default: http://localhost:11434) and select your model
4. AI features (health analysis, RAG, embeddings) will use the configured model

• AES-256-GCM encryption for sensitive data
• Credential vault with Argon2id key derivation
• Secure memory handling for cryptographic operations
• No background telemetry
• All data stored locally

MIT License - See LICENSE

See CONTRIBUTING.md