AISSIA - AI Smart Schedule & Interactive Assistant

AISSIA is an intelligent personal assistant for time management, hyperfocus detection, and language learning, powered by GroveEngine.

What is AISSIA?

AISSIA helps manage:

• Hyperfocus: Detects when you've been working too long and need a break
• Time Management: Intelligent scheduling and task planning
• Language Learning: Interactive practice in target languages
• Notifications: Context-aware alerts with TTS support

Built on GroveEngine

AISSIA leverages GroveEngine, a C++17 hot-reload module system that enables:

• Hot-Reload: Modify code at runtime without losing state (<1ms reload latency)
• Modular Architecture: Self-contained modules (200-300 lines each)
• Pub/Sub Communication: Decoupled inter-module messaging
• State Preservation: Automatic state serialization across reloads
• Configuration Hot-Reload: Update settings without code changes

Why GroveEngine?

✅ Ultra-fast Development: Hot-reload validated at 0.4ms ✅ Type Safety: Strong C++ typing, no "wildcode" ✅ Proven Architecture: Production-ready module system ✅ Privacy-First: Local mode, data never uploaded ✅ Progressive Evolution: MVP → Production → Cloud without rewrite

Project Structure

Aissia/
├── external/
│   └── GroveEngine/          # GroveEngine (symlink)
├── src/
│   ├── main.cpp              # Main application loop
│   └── modules/
│       ├── SchedulerModule.*       # Time management & hyperfocus
│       ├── NotificationModule.*    # Alerts & TTS
│       ├── AIAssistantModule.*     # (TODO) LLM integration
│       ├── LanguageLearningModule.* # (TODO) Language practice
│       └── DataModule.*            # (TODO) SQLite persistence
├── config/
│   ├── scheduler.json
│   └── notification.json
├── docs/
│   ├── README.md                   # Project documentation
│   └── GROVEENGINE_GUIDE.md        # GroveEngine user guide
└── CMakeLists.txt

Modules

Implemented

SchedulerModule: Time management and hyperfocus detection

• Tracks work sessions and break intervals
• Detects hyperfocus based on configurable threshold (default: 120 minutes)
• Reminds for breaks (default: every 45 minutes)
• Task estimation and tracking

NotificationModule: System alerts with priority levels

• Priority-based notifications (LOW, NORMAL, HIGH, URGENT)
• Silent mode (respects URGENT priority)
• TTS support (configurable)
• Multilingual support (fr/en/jp)
• Queue management with rate limiting

Planned

AIAssistantModule: LLM-powered contextual interventions

• Claude API integration
• Context-aware suggestions
• Intelligent document retrieval

LanguageLearningModule: Language practice and learning

• Conversation in target language
• Vocabulary tracking
• Progress monitoring

DataModule: SQLite persistence

• Task history
• Session analytics
• Configuration backup

Getting Started

Prerequisites

• CMake 3.20+
• C++17 compiler (GCC/Clang)
• GroveEngine (included via symlink)

Build

# Configure
cmake -B build

# Build everything
cmake --build build -j4

# Build modules only (for hot-reload workflow)
cmake --build build --target modules

Run

# Interactive mode (requires ANTHROPIC_API_KEY in .env)
./run.sh

# Or manually with .env loaded
source .env && ./build/aissia -i

# MCP Server mode (for Claude Code integration)
./build/aissia --mcp-server

# Normal mode (background daemon)
./build/aissia

Note: Create a .env file with your API key:

ANTHROPIC_API_KEY=sk-ant-...

Hot-Reload Workflow

1. Start AISSIA: ./build/aissia
2. Edit a module: src/modules/SchedulerModule.cpp
3. Rebuild: cmake --build build --target modules
4. Module reloads automatically with state preserved

Configuration

Configuration files in config/:

scheduler.json:

{
    "hyperfocusThresholdMinutes": 120,
    "breakReminderIntervalMinutes": 45,
    "breakDurationMinutes": 10,
    "workdayStartHour": 9,
    "workdayEndHour": 18
}

notification.json:

{
    "language": "fr",
    "silentMode": false,
    "ttsEnabled": false,
    "maxQueueSize": 50
}

Documentation

• docs/README.md: Project documentation and architecture
• docs/GROVEENGINE_GUIDE.md: Complete GroveEngine user guide
• CDCDraft.md: Detailed requirements specification (French)

Development Philosophy

MVP-First

• Phase 1 (Required): Local Windows mode only
• Phase 2+ (Optional): Cloud features if needed
• Configuration-Driven: 90% of needs via JSON
• Simplicity First: Complexity emerges from interaction

Module Development

Each module is a self-contained unit (~200-300 lines):

1. Implements IModule interface
2. Pure business logic (no infrastructure code)
3. Communicates via IIO pub/sub
4. Serializes state for hot-reload
5. Configurable via IDataNode

See docs/GROVEENGINE_GUIDE.md for detailed module development guide.

Architecture

AISSIA runs on a simple loop (10Hz for assistant workload):

┌─────────────────────────────────────────────┐
│           Main Loop (10Hz)                   │
│                                              │
│  ┌────────────────────────────────────┐    │
│  │  1. Check for module file changes   │    │
│  │  2. Hot-reload if modified          │    │
│  │  3. Process all modules             │    │
│  │  4. Route inter-module messages     │    │
│  │  5. Sleep to maintain 10Hz          │    │
│  └────────────────────────────────────┘    │
└─────────────────────────────────────────────┘

Modules communicate via topics:

• scheduler:hyperfocus_detected → NotificationModule alerts user
• notification:alert_sent → DataModule logs event
• aiassistant:suggestion → NotificationModule displays suggestion

Technical Stack

• Core: C++17
• Module System: GroveEngine
• Logging: spdlog
• Data Format: JSON (nlohmann/json)
• Build: CMake 3.20+
• Future: SQLite (DataModule), Claude API (AIAssistantModule)

Roadmap

Completed ✅

• Project setup with GroveEngine
• SchedulerModule (time management & hyperfocus)
• NotificationModule (alerts & TTS)
• MonitoringModule (app tracking & classification)
• AIModule (LLM conversation & suggestions)
• VoiceModule (TTS/STT interfaces)
• StorageModule (SQLite persistence)
• WebModule (HTTP requests via IIO pub/sub)
• LLM Service (Claude API integration - Sonnet 4)
• 17 Agentic Tools (filesystem, scheduler, storage, etc.)
• MCP Server mode (JSON-RPC stdio)
• Interactive mode (--interactive)
• 110/110 tests passing

In Progress 🚧

• Test interactive mode with real queries
• Expose InternalTools via MCP Server

Planned 📋

• Windows Toast notifications
• Real TTS integration (espeak-ng)
• Language learning features

License

To be determined

Links

• GroveEngine: ../GroveEngine
• Claude Code: https://docs.claude.com/en/docs/claude-code