
Foil uses on-device LLMs to scan your code for security flaws. Your source code never leaves your Mac. Deep Dive & Deep Scan explain the vulnerability and rewrite the fix.
Trusted by developers at
We just released Foil and we're building it up. No fake logos here. If you'd like to help shape it, come join the beta.
Step through a real scan — click through the dashboard, Deep Dive & Deep Scan, and rewritten fixes in this interactive walkthrough.
Foil ships with SecureReview-7B — our own LLM fine-tuned specifically for code review and vulnerability analysis, quantized to 4-bit MLX for native Apple Silicon performance.
While competitors rely on generic cloud models that guess at security context, SecureReview-7B was trained on the agentic workflow Foil uses to validate findings, explain impact, and rewrite fixes.
SecureReview-7B
4-bit MLX · Apple Silicon
Parameters
7B
Quantization
4-bit
Framework
MLX
Context
8K tokens
Traditional scanners match patterns. Foil reasons about your code. Click a card with a play icon to watch it in action.
According to the OWASP Top 10 (2025), broken access control remains the #1 web application security risk. The IBM Cost of a Data Breach Report 2024 found the average breach costs $4.88M.
Detailed feature comparisons with popular security tools.
Foil vs Semgrep
Pattern matching vs AI reasoning
Foil finds logic flaws Semgrep can't→
Foil vs Snyk Code
Cloud analysis vs 100% local
Your code never leaves your machine→
Foil vs CodeQL
Query language vs instant scanning
No query language to learn→
Foil vs Foil
Which model & flags to pick for your case
See benchmarks & pick a preset→
$ brew tap vitorallo/foil
$ brew install --cask foil
$ foil scan ./my-project
✓ Scanning 47 files...
⚠ 3 vulnerabilities found
$ foil scan --deep ./my-project
🔍 Analyzing findings with local LLM...
✓ 3 fixes generated with explanations
Latest posts on security, AI, and developer tooling.