Solo dev, no VC, one mission: make AI compliance scannable before August 2026
Hey PH community. I'm Jason, a solo founder building AIR Blackbox, an open-source CLI that scans Python AI projects for EU AI Act compliance.
Quick background: I spent 13+ years as a technical sourcer at Meta and AWS, where I watched companies move fast on AI adoption and slow on governance. When the EU AI Act passed with an August 2, 2026 enforcement deadline, I realized most teams had no idea what the technical requirements actually look like in code.
So I built a linter for AI governance.
pip install air-blackbox
air-blackbox discover .
10 seconds. 14+ checks mapped to the six EU AI Act articles (risk management, data governance, documentation, record-keeping, human oversight, robustness). Runs locally, no API keys, no cloud dependency.
The part I'm most proud of: I scanned real open-source frameworks and published the results publicly. Haystack's engineering team lead responded to the GitHub issue within 38 minutes and confirmed the findings. That kind of open, verifiable accountability is the whole point.
The tool also provides runtime validation rules (tool allowlists, PII detection, hallucination guards, content policy enforcement) as drop-in trust layers for LangChain, CrewAI, AutoGen, and other frameworks.
I'm a solo dev on this. No VC, no team, just shipping. Would love to connect with other builders working in AI governance, compliance tooling, or open-source dev tools.
Curious: for those of you building AI products, has the EU AI Act even hit your radar yet? Or is it still something you figure you'll deal with later?
GitHub: github.com/air-blackbox
Website: airblackbox.ai
Replies