EU ACT Guard
EU AI Act compliance scanner that reads your actual code
Start new thread
Dhanmeet Nijhawan

EU ACT Guard - EU AI Act compliance scanner that reads your actual code

by•
Every EU AI Act tool asks questions. EU ACT Guard reads your actual code from GitHub. We detect violations with exact file paths and fixes — not generic checklists. If there’s an Article 12 logging issue, we show the missing line and the code to patch it. One scan covers: → GitHub repo compliance → Website crawl for Article 50 labeling → Privacy policy checks → Annex IV draft → ISO 42001 mapping Three reports: Technical, Legal, Business. EU AI Act enforcement starts Aug 2, 2026.

Add a comment

Replies

Best
Dhanmeet Nijhawan
Maker
📌
Hey Product Hunt 👋 I built EU ACT Guard after watching a compliance consultant spend 4 weeks asking a client questions about their AI system — never once looking at the actual code. The EU AI Act enforcement deadline is August 2, 2026. Most tools help companies document what they claim to do. We check what their code actually does. The difference matters: your questionnaire says "we have human oversight." Your code has AI writing directly to your database with no review step. That is an Article 14 violation — and a regulator will find it even if your compliance document says otherwise. Our scanner caught this pattern in 6 of the first 10 repos we tested publicly. What makes us different: - File paths + line numbers for every violation - Copy-paste code fix for each finding - Contradiction detection — questionnaire vs actual code - Annex IV technical file generated automatically - Three report formats (developer / lawyer / founder) - Covers EU AI Act + ISO 42001 + prEN 18286 We are in early access — free scan for everyone who signs up today. Would love your feedback on what compliance checks matter most to your team. Happy to answer anything 🙏
Danush Singla

Interesting launch, Dhanmeet. I’m not based in the EU, so I don’t have much direct skin in the game here, but I’ve seen a few AI Act tools and the code-scanning angle is definitely different.

The part that caught my attention is the gap between what a company says in a compliance questionnaire and what the code actually does. For teams trying this, how are you thinking about the trust step before they connect a repo? Is the main hurdle explaining the legal value of the scan, or making developers comfortable with code access first?