A private, security-first Alfred-style AI assistant for Linux that combines local or optional API-based language models with a strictly controlled command execution system. It runs offline by default, supports GPU-accelerated local models, and enf…
A private, security-first Alfred-style AI assistant for Linux that combines local or optional API-based language models with a strictly controlled command execution system. It runs offline by default, supports GPU-accelerated local models, and enforces deterministic, whitelist-based actions so the AI can generate text without ever directly accessing or controlling the system.
Used Vibe-coding for most of the part in this project. But worked on the debugging myself.