Foil AI Code Security

Foil runs an AI‑powered security scan entirely on a Mac, using an on‑device large language model to analyze source code for vulnerabilities without sending any data to the cloud. The native macOS application and a CLI command can examine projects written in several programming languages, producing detailed explanations of identified issues and automatically generated fixes.

The tool includes a custom‑trained 7‑billion‑parameter model, SecureReview‑7B, quantized to 4‑bit for efficient execution on Apple‑Silicon GPUs. This model is fine‑tuned for code review and vulnerability reasoning, enabling it to reason about security flaws rather than merely matching patterns. Results are presented through “Deep Dive” and “Deep Scan” views that validate exploitability and suggest rewritten code.

Because the analysis runs locally, developers retain full control over their code and avoid any external data exposure. The product is currently in public beta and positioned as a free community edition with optional paid plans for additional features.