@meok-ai-labs/mcp-eu-ai-act

MCP (Model Context Protocol) server for EU AI Act compliance checking and documentation generation.

Installation

npm install -g @meok-ai-labs/mcp-eu-ai-act

Usage with Claude Desktop

Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "eu-ai-act": {
      "command": "npx",
      "args": ["-y", "@meok-ai-labs/mcp-eu-ai-act"]
    }
  }
}

Available Tools

classify_ai_risk

Classify an AI system under EU AI Act risk categories.

Input:

• use_case: The AI use case (e.g., "credit scoring", "hiring decision")

Output: Risk classification (PROHIBITED / HIGH_RISK / LIMITED_RISK), applicable article, penalties, and required actions.

check_compliance

Run a full compliance check against all 5 EU AI Act pillars.

Input:

• system_name: Name of the AI system
• use_case: The AI use case

Output: Compliance score (0-100%), pass/fail status, and specific gaps with recommendations.

generate_compliance_report

Generate a complete compliance report with risk classification, gap analysis, and remediation roadmap.

Input:

• system_name: Name of the AI system
• company: Company name
• use_case: The AI use case

Output: Full formatted report ready for regulators or board presentation.

get_deadline_info

Get current EU AI Act enforcement deadline and days remaining.

Output: Deadline date, days remaining, penalty information.

Example Conversation with Claude

You: Check if our credit scoring system is compliant

Claude: I'll run a compliance check for your credit scoring system. [Uses classify_ai_risk → HIGH_RISK, Article 6] [Uses check_compliance → 32% score, 17 gaps identified]

Your credit scoring system is classified as HIGH-RISK under EU AI Act Article 6. The compliance score is 32% with 17 critical gaps identified...

Compliance Pillars Checked

1. Data Governance (Article 10) — Training data quality, bias mitigation, GDPR compliance
2. Technical Documentation (Article 11) — Model cards, architecture docs, version control
3. Transparency (Article 13) — User disclosure, explainability, confidence scores
4. Human Oversight (Article 14) — Intervention protocols, escalation procedures
5. Accuracy & Robustness (Article 15) — Testing, error rates, fail-safes

About MEOK Labs

Built by MEOK AI Labs — sovereign AI for British critical infrastructure.

• Website: https://meok.ai
• Email: [email protected]
• EU AI Act audits: https://meok.ai/audit

License

MIT