Free tool · EU AI Act
EU AI Act Agent Compliance Scanner
Scan your AI agent deployment against EU AI Act requirements and generate a compliance checklist with risk scoring. Everything runs in your browser; nothing you enter leaves your device.
Introduction
AI agents are not ordinary software features. An agent that reads data, calls tools, and acts on its own sits squarely inside the EU AI Act's risk-based duties — risk management, data governance, technical documentation, record-keeping, transparency, human oversight, and post-market monitoring. This scanner turns those duties into a concrete checklist for one deployment: describe the agent, tick the controls you actually have, and get a risk score plus the exact gaps to close, in priority order.
Compliance Scanner
Regulatory requirements — tick what is implemented and tested
Tick every requirement that is actually in place for this deployment — not planned, not partial.
Results and Recommendations — your compliance checklist
This is an informational self-assessment, not legal advice. It maps operational controls against EU AI Act requirements — it does not determine your legal classification under the EU AI Act. For binding answers, consult qualified counsel.
Results and Recommendations
Your score weights each missing requirement by how much it matters for your risk profile: a missing audit trail or missing human oversight on a high-risk, autonomous agent scores worse than the same gap on a minimal-risk drafting tool. Work the checklist top-down — critical gaps first — and re-run the scan whenever the agent, its autonomy, its data, or its use case changes. Compliance decays silently as agents gain new tools and new data access.
- Risk management: a documented process for identifying and mitigating the agent's foreseeable risks, kept current across its lifecycle.
- Data governance: documented sources, rights, lineage, and retention for the data the agent reads and learns from.
- Technical documentation: architecture, intended use, known limits, and evaluation results, ready for a regulator or enterprise buyer.
- Record-keeping: timestamped logs of agent actions, approvals, overrides, and stops, under a retention policy.
- Transparency: people interacting with the agent — or affected by its output — know it, and know when a human can step in.
- Human oversight: named overseers who understand the system, can interpret its output, and can override or stop it.
FAQ
Is this legal advice?
No. This tool is an informational self-assessment for operational readiness — it is not legal advice and does not determine your legal classification under the EU AI Act.
Does anything I enter leave my device?
No. The scan runs entirely in your browser with JavaScript. Nothing you enter is sent to a server.
What is the buyer keyword this tool answers?
EU AI Act compliance checklist — the scan generates exactly that: a prioritised compliance checklist with risk scoring for an AI agent deployment.
How is risk scoring calculated?
Each unticked requirement counts as a gap, weighted by requirement importance and by your use-case, autonomy, and data selections. The score is the weighted share of requirements in place, out of 100.
Close the gaps, don't just list them
Get the $29 Governed Agent Mesh Playbook to ensure regulatory readiness — the approval-gate patterns, audit-trail templates, oversight procedures, and incident runbook behind every fix on your checklist, ready to run.
Get the Studio Edition — $29