AI-NAS: Intelligent Storage Management

AI-NAS is a smart Network Attached Storage (NAS) management solution that integrates a powerful AI Assistant to simplify data management, system monitoring, and file operations through natural language.

• Conversational AI Assistant: Manage your NAS using natural language commands.
• Real-time Streaming: Instant feedback for AI responses using server-sent events or streaming repositories.
• Context-Aware File Operations: Attach files directly from your NAS to the AI chat for analysis, summarization, or searching.
• Quick Action Chips: One-tap shortcuts for common tasks:
  • Storage Health: "How much storage is left?"
  • Deep Search: "Find all PDF files in /Home"
  • System Optimization: "Run a performance check"
• Versatile AI Assistant:
  • Explain images
  • Lables on images
  • ...
• RAG with elasticsearch: Exact retrieval and semantic similarity retrieval.
• Cross-Platform GUI: Responsive frontend built with Flutter.

• Quant models with llama.cpp to run on resource constrained devices.

• /frontend: The Flutter-based client application.
• /vendor: Custom Flutter toolchain and engine configurations.

• Frontend: Flutter / Dart
• State Management: standard StatefulWidget patterns (extensible to Riverpod/Bloc).
• Communication: REST API with real-time stream support.
• Localization: Official Flutter i18n support (ARB files).

1. Prerequisites: Ensure you have the Flutter SDK installed.
2. Installation:

   bash bootstrap.sh --setup

3. Run Backend:

   bash bootstrap.sh --backend

4. Run Frontend:

   bash bootstrap.sh --frontend

Pre-built images are available on GitHub Container Registry:

Run the CPU variant:

docker run -d \
  --name ainas \
  -p 9026:9026 \
  -v ./storage:/app/storage \
  ghcr.io/zegang/ainas-backend:cpu

For NVIDIA GPU:

docker run -d \
  --name ainas \
  --gpus all \
  -p 9026:9026 \
  -v ./storage:/app/storage \
  ghcr.io/zegang/ainas-backend:cuda

For AMD GPU:

docker run -d \
  --name ainas \
  --device=/dev/kfd --device=/dev/dri \
  --group-add video \
  -p 9026:9026 \
  -v ./storage:/app/storage \
  ghcr.io/zegang/ainas-backend:rocm

The container serves both the backend API and the Flutter web frontend (at http://localhost:9026).

bash bootstrap.sh --build-backend-image cpu     # CPU (default)
bash bootstrap.sh --build-backend-image cuda    # NVIDIA GPU
bash bootstrap.sh --build-backend-image rocm    # AMD GPU