██████╗  █████╗ ██╗███████╗███████╗
                              ██╔══██╗██╔══██╗██║██╔════╝██╔════╝
                              ██████╔╝███████║██║███████╗███████╗
                              ██╔══██╗██╔══██║██║╚════██║╚════██║
                              ██████╔╝██║  ██║██║███████║███████║
                              ╚═════╝ ╚═╝  ╚═╝╚═╝╚══════╝╚══════╝

Choose the appropriate version for your system:

macOS

• Download for Apple Silicon (M1/M2/M3/M4) - ARM64
• Download for Intel Mac - x64

Windows

• Download for Windows - x64

  Note: Windows SmartScreen may show a warning as we're awaiting code signing approval. The app is safe to install.

Not sure which version?

• Mac users: If you bought your Mac in 2020 or later, choose Apple Silicon. Otherwise, choose Intel.
• Windows users: Most modern Windows PCs use x64 architecture.

View all releases

Note: Beta releases are experimental builds from our dev branch. We don't recommend using beta versions unless you're a developer who wants to test new features or debug issues. Beta releases may be unstable and contain bugs.

In an era where data privacy is paramount, Baiss brings the power of Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) directly to your desktop—running entirely locally.

No cloud subscriptions, no data leaks, just pure AI productivity. Whether you're a developer needing a coding assistant or a researcher organizing documents, Baiss provides a unified, cross-platform interface to interact with your data and models.

• Privacy First: Runs local LLMs (via llama.cpp) and vector search on your machine. Your data never leaves your device.
• Advanced RAG: Built-in Retrieval-Augmented Generation using DuckDB for high-performance vector storage and FlashRank for re-ranking.
• Cross-Platform UI: A beautiful, responsive interface built with Avalonia UI, running natively on macOS, Windows, and Linux.
• Extensible Architecture: Designed with Clean Architecture principles, making it easy for developers to add new AI providers, tools, or plugins.
• Python Power: Leverages a robust Python backend (FastAPI) for heavy AI lifting, seamlessly integrated with the .NET frontend.

Frontend & Core:

• C# / .NET 8: The backbone of the application.
• Avalonia UI: For a pixel-perfect cross-platform user experience.
• Clean Architecture: Separation of concerns (Domain, Application, Infrastructure, UI).

AI Backend:

• Python & FastAPI: Handles AI logic and API endpoints.
• DuckDB: Embedded SQL OLAP database for efficient vector search.
• Llama.cpp: For running quantized LLMs locally with hardware acceleration.
• HuggingFace & Transformers: For embeddings and model management.

Ensure you have the following installed:

• .NET 8 SDK: Download here
• Python 3.10+: Download here

1. Clone the repository:

   git clone https://github.com/Tbeninnovation/Baiss.git
   cd Baiss

2. Set up the Python Environment: Navigate to the core directory and install dependencies.

   cd core/baiss
   pip install -r requirements.txt

To run the application locally:

# Navigate to the UI project
cd Baiss.UI

# Run the application
dotnet run

Note: On the first run, Baiss may need to download default models or configure the local database. Please check the console output for status updates.

Here's a quick look at the codebase organization:

Baiss/
├── Baiss.UI/              # Avalonia UI Frontend & Entry Point
├── Baiss.Application/     # Business Logic, Interfaces, Use Cases
├── Baiss.Domain/          # Core Entities & Value Objects
├── Baiss.Infrastructure/  # Services, DB Access, External APIs
└── core/
    └── baiss/             # Python Backend (FastAPI, Agents, RAG)
        ├── requirements.txt
        └── shared/python/baiss_agents/

We love contributions! Whether it's fixing a bug, improving the UI, or adding support for a new AI model, your help is welcome.

1. clone the repo and create a branch from dev
2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
3. Commit your Changes (git commit -m 'Add some AmazingFeature')
4. Push to the Branch (git push origin feature/AmazingFeature)
5. Open a Pull Request to dev
6. We will test in dev and merge to main when ready.

A huge thank you to the brilliant minds building Baiss:

• @taharbmn - Thought this was a 2-week project. That was 6 months ago.
• @Abdelmathin - The voice of reason we muted on Meetings.
• @L0Abdellah - The only one who knows why the search results actually work.
• @AYoubZarda - Burned a laptop developing this (RIP)
• @DraGSsine - Laptop Survived, My Code Didn’t