Installation

System requirements and how to install ForgeAI

System Requirements

Minimum

• OS: Linux, macOS, or Windows
• RAM: 8 GB
• Disk: 2 GB + space for your models
• CPU: x86_64 or ARM64

Recommended

• RAM: 16 GB+
• GPU: NVIDIA (CUDA), AMD (Vulkan), or Apple Silicon (Metal)
• Disk: SSD with 50 GB+ free
• Python 3.10+ for Training & Convert

Install from Release

1. Download the release

   Download the latest release for your platform from the releases page.

   Platform	File
   Linux	forgeai_x.x.x_amd64.AppImage or .deb
   macOS	ForgeAI_x.x.x.dmg
   Windows	ForgeAI_x.x.x_x64-setup.exe

2. Run the installer

   • Linux: Make the AppImage executable (chmod +x) and run, or install the .deb
   • macOS: Open the .dmg and drag ForgeAI to Applications
   • Windows: Run the setup executable
3. First launch

   Open ForgeAI. You’ll land on the Dashboard (module 00).

Build from Source

1. Install prerequisites

   • Rust (latest stable)
   • Node.js (v20+)
   • Tauri v2 prerequisites for your OS
2. Clone and install

   git clone https://github.com/siddhesh2377/ForgeAi.git
   cd ForgeAi
   npm install
   
3. Run in development mode

   npm run tauri dev
   
4. Build for production

   npm run tauri build
   

   The compiled binary will be in src-tauri/target/release/.

Optional: GPU Setup

For accelerated inference and training:

NVIDIA (CUDA)

1. Install NVIDIA drivers (515+ recommended)
2. CUDA toolkit is bundled with llama.cpp — no separate install needed
3. In ForgeAI Settings, install llama.cpp with the CUDA variant
4. For training, CUDA 11.8+ is required (PyTorch auto-detects during setup)

AMD / Intel (Vulkan)

1. Install Vulkan drivers for your GPU
2. In ForgeAI Settings, install llama.cpp with the Vulkan variant
3. Training is CPU-only on AMD/Intel (PyTorch CUDA not available)

Apple Silicon (Metal)

1. Metal is built into macOS — no extra drivers needed
2. In ForgeAI Settings, install llama.cpp with the default variant
3. Training uses MPS (Metal Performance Shaders) automatically

CPU Only

No extra setup needed. In ForgeAI Settings, install llama.cpp with the CPU variant.

Optional: Python (for Training & Convert)

Both the Training module and Convert module require Python 3.10+. Each uses its own isolated virtual environment.

1. Verify Python

   python3 --version
   

   If not installed, get it from python.org or your package manager.

2. Auto-setup in ForgeAI

   Navigate to the Training or Convert module. ForgeAI will detect Python and prompt you to install dependencies. Each environment is a one-time setup:

   Environment	Size	Packages
   Training	~2 GB	PyTorch, Transformers, PEFT, TRL, BitsAndBytes
   Convert	~500 MB	Transformers, Torch, SafeTensors, SentencePiece

   You can also manage both environments from Settings (07).