Silicon Studio
The All-in-One Local AI Workbench for Apple Silicon
Silicon Studio is a powerful, open-source desktop application designed to make local LLM fine-tuning and inference accessible to everyone on Apple Silicon Macs. Built on top of Apple’s MLX framework, it provides a unified interface for data preparation, model management, fine-tuning, and evaluation.
✨ Features
🍎 Native Apple Silicon Performance
Leverage the full power of your M1/M2/M3/M4 chips. Silicon Studio uses the MLX framework for efficient, hardware-accelerated fine-tuning (LoRA/QLoRA) and inference directly on your Mac.
📊 Data Preparation Studio
A built-in studio to clean, format, and prepare your training datasets.
- Preview & Edit: V…
Silicon Studio
The All-in-One Local AI Workbench for Apple Silicon
Silicon Studio is a powerful, open-source desktop application designed to make local LLM fine-tuning and inference accessible to everyone on Apple Silicon Macs. Built on top of Apple’s MLX framework, it provides a unified interface for data preparation, model management, fine-tuning, and evaluation.
✨ Features
🍎 Native Apple Silicon Performance
Leverage the full power of your M1/M2/M3/M4 chips. Silicon Studio uses the MLX framework for efficient, hardware-accelerated fine-tuning (LoRA/QLoRA) and inference directly on your Mac.
📊 Data Preparation Studio
A built-in studio to clean, format, and prepare your training datasets.
- Preview & Edit: Visualize JSONL and CSV datasets instantly.
- Privacy Shield: Automatically strip PII (Personally Identifiable Information) from your datasets using local NLP models.
- Format Conversion: Convert between common formats (CSV -> JSONL) compatible with Llama, Mistral, and other chat templates.
🤖 Model Management
Browse, download, and manage open-source models directly from Hugging Face.
- One-Click Downloads: Supports popular architectures like Llama 3, Mistral, Qwen 2.5, Gemma, and Phi.
- Quantization Support: Run large models efficiently with 4-bit and 8-bit quantization.
🦾 Fine-Tuning Engine
Fine-tune models on your own data without a cloud GPU.
- LoRA / QLoRA: Efficient low-rank adaptation for fine-tuning large models with minimal VRAM.
- Visual Configuration: Configure learning rates, epochs, rank, and more with a simple UI.
- Real-time Monitoring: Watch loss curves and training metrics in real-time.
💬 Local Chat Inference
Test your fine-tuned models immediately in a ChatGPT-like interface.
- Private & Offline: Your data never leaves your device.
- Instant Switching: Toggle between base models and your fine-tuned adapters.
🚀 Getting Started
Installation
Option 1: Download Release (Recommended) Download the latest .dmg from the Releases page.
Option 2: Build from Source
Prerequisites:
- macOS 13+ (Ventura or later)
- Apple Silicon Mac
- Node.js 18+
- Python 3.10+ (Anaconda recommended)
# Clone the repository
git clone [https://github.com/yourusername/silicon-studio.git](https://github.com/rileycleavenger/Silicon-Studio.git)
cd silicon-studio
# Install dependencies
npm install
# Setup Backend Environment
cd backend
# Create a conda env (recommended) or venv
conda create -n silicon-studio python=3.11
conda activate silicon-studio
pip install -r requirements.txt
cd ..
# Run Development Mode
npm run dev
🛠️ Tech Stack
- Frontend: Electron, React, TypeScript, Vite, TailwindCSS
- Backend: Python, FastAPI
- AI Engine: Apple MLX, MLX-LM
- Data Processing: Pandas, Presidio (Microsoft), Spacy
🤝 Contributing
Contributions are welcome! We want to make local AI accessible to everyone. Refactoring, documentation, and new features are all appreciated.
- Fork the Project
- Create your Feature Branch (
git checkout -b feature/AmazingFeature) - Commit your Changes (
git commit -m 'Add some AmazingFeature') - Push to the Branch (
git push origin feature/AmazingFeature) - Open a Pull Request
📄 License
Distributed under the MIT License. See LICENSE for more information.
Built with ❤️ for the Local AI Community