The Local AI Playground

The Local AI Playground is a native app for running and managing AI models locally, without complex setup and without needing a GPU. It’s designed for experimenting on your own computer while keeping data private and maintaining full control of the environment.

Local runs and privacy

All computation happens on your device, so prompts and responses don’t go to the cloud. This is useful for developers, companies, and researchers working with sensitive information.

Performance on CPU

The app uses a Rust backend to stay lightweight and efficient on CPU. It can adapt to available CPU threads and supports GGML quantization options, including:

• q4
• 5.1
• 8
• f16

This makes it possible to run models such as WizardLM 7B in just a couple of clicks.

Cross-platform and open source

The Local AI Playground is available on:

• Windows (.MSI, .EXE)
• macOS (M1/M2, Intel)
• Linux (AppImage, .deb)

The project is free and open source. Planned improvements include GPU support and tighter integration with external AI apps, for example via window.ai.