eu-ai-act-compliance-mcp 1.4.1

EU AI Act compliance for AI agents. 410 articles from EUR-Lex via FTS5 search. Instant risk scan, deadline tracker, 42-point audit, documentation generator, penalty calculator. Zero-config free tier. Built by MEOK AI Labs.

EU AI Act + Multi-Regulation Compliance MCP Server

Six EU regulations, one MCP. Verbatim text + active compliance scanning + cryptographic attestations.

Covers EU AI Act · DORA · NIS2 · Cyber Resilience Act · CSRD · GDPR — 400+ articles indexed for FTS5 search, every quote auditor-defensible, every citation linked back to the canonical EUR-Lex URL.

🆕 v1.4 — Verbatim EU regulation text from publications.europa.eu Cellar SPARQL, in SQLite FTS5. Daily sync.

Search regulation text · Quote full articles · Classify AI risk levels · Run 42-point audits · Generate Annex IV docs · Assess penalties · Track deadlines · Sign attestations

Installation · Tools · Docs · Report Bug

---

Quick Install

Client	Install
Claude Desktop
Cursor
VS Code
Windsurf
Docker	docker run -p 8000:8000 eu-ai-act-compliance-mcp
pip	pip install eu-ai-act-compliance-mcp

Connect via MCPize

Use this MCP server instantly with no local installation:

npx -y mcpize connect @meok-ai-labs/eu-ai-act-compliance --client claude

Or connect at: https://mcpize.com/mcp/eu-ai-act-compliance

---

Quick Start

pip install eu-ai-act-compliance-mcp
# or
npm install -g @meok-ai/eu-ai-act-compliance-mcp

Why This Exists

The EU AI Act (Reg 2024/1689) is now in force. Following the March 2026 Digital Omnibus vote, the timeline shifted:

• Article 50 transparency obligations: 2 November 2026 (was August 2026)
• Annex III high-risk systems: 2 December 2027 (was August 2026)
• Annex I high-risk systems: 2 August 2028 (was August 2027)

Penalties remain unchanged: up to €35M or 7% of global turnover.

Most teams are using PDF binders and Word checklists to track Article 6 risk classifications, Article 26(9) FRIA artifacts, and Article 50 disclosures. When a regulator asks "how do we know this artifact wasn't fabricated last week?", the answer today is "trust us".

This MCP turns Article 6 / 26(9) / 50 obligations into a single AI-agent-callable tool, signs each artifact with HMAC-SHA256, and gives you a verifiable URL the auditor can curl independently.

Real Usage Example

A German Mittelstand HR-tech firm needed to dry-run their Article 6 classification + Article 26(9) FRIA for a CV-screening AI. Their compliance lead installed this MCP into Claude Code:

pip install eu-ai-act-compliance-mcp

Then prompted Claude:

"Classify our CV-scoring product against EU AI Act Article 6. Treat it as Annex III (employment). Generate the risk-tier rationale and the high-risk obligations checklist. Then produce the Article 26(9) FRIA. Sign with the attestation API."

Result: 49-page audit pack with cryptographically verifiable HMAC-signed sections in ~14 hours of review time.

Traditional consulting estimate: 230 hours / £42-62K.

Saved: ~£40K and 4-5 weeks.

Tools

🆕 v1.4 — EUR-Lex Search (free tier)

Tool	Description
search_regulation	Full-text FTS5 search across verbatim EU regulation text (EU AI Act, DORA, NIS2, CRA, CSRD, GDPR). Returns 64-token snippets with relevance scores and a canonical EUR-Lex URL for every hit.
get_article_text	Return the full verbatim text of a single article (e.g. GDPR Article 33) plus its canonical EUR-Lex URL. Drop straight into audit evidence packs.
list_regulations_in_db	List all regulations in the local DB with article counts + last-sync date.

Core compliance tools

Tool	Description
quick_scan	One-sentence AI system description → instant risk classification (no API key)
deadline_check	All EU AI Act enforcement deadlines with days remaining (zero params)
classify_ai_risk	Detailed Article 5/6/50 risk classification
check_compliance	42-point compliance audit against Annex I-IX
generate_annex_iv_docs	Article 11 technical documentation generator
assess_penalties	Penalty exposure calculator (up to €35M or 7% global turnover)
multi_jurisdiction_map	Cross-border compliance mapping
predict_risk_neural	Neural-net risk prediction (Pro tier)
neural_insights	Compliance pattern insights from training data (Pro tier)

Example: search the EU AI Act for "biometric"

result = search_regulation(query="biometric", regulation="eu-ai-act", limit=3)

Returns matched snippets from Article 3 (definitions), Article 5 (prohibitions), Article 26 (deployer duties), with relevance scores and >>>highlight<<< markers.

Why FTS5?

• Verbatim text — no LLM summarization, every quote is auditor-defensible
• Token-safe — 64-token snippets fit in any context window
• Daily sync — GitHub Actions polls EUR-Lex Atom feed at 06:00 UTC
• Stdlib only — no Postgres, no external deps

Pricing

Tier	Price	What you get
Free	£0/forever	10 calls/day — quick_scan, deadline_check, risk classification (summary)
Starter	£29/mo	100 calls/day — full detailed analysis + Annex IV docs + audit reports
Professional	£79/mo	1,000 calls/day — multi-jurisdiction mapping + neural predictions + attestations
Enterprise	Custom	Unlimited — on-premise + custom models + SLA + SSO

Get your API key →

---

If this tool helps your compliance workflow, please star this repo — it helps other compliance teams find it and keeps it maintained.

→ Subscribe to Pro · Enterprise · Book assessment

Attestation API

Every Pro/Enterprise audit produces a cryptographically signed certificate:

POST https://meok-attestation-api.vercel.app/sign
→ { cert_id, verify_url, hmac_sha256, valid_until }

Verify any certificate: https://meok-attestation-api.vercel.app/verify/{cert_id}

Or install the zero-dep verifier: pip install meok-attestation-verify

Star History

Need Full EU AI Act Compliance?

This MCP gives you the tools — councilof.ai gives you the full platform.

Tier	Price	What You Get
Starter	£29/mo	Automated risk classification + deadline tracking
Pro	£79/mo	Full audit packs + HMAC-signed attestations
Enterprise	£1,499/mo	Dedicated compliance support + Notified Body prep
Gap Analysis	£5,000	48-hour expert assessment with signed report

→ Get started at councilof.ai — 100x cheaper than traditional compliance consulting.

Support & Enterprise

• GitHub Discussions
• Report Issues
• Enterprise support: [e-mail]
• Website: meok.ai
• All MCP servers: meok.ai/labs/mcp/servers
• Attestation API: meok-attestation-api.vercel.app
• Compliance platform: councilof.ai

License

MIT © MEOK AI Labs

<script type="application/ld+json"> { "@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [ { "@type": "Question", "name": "Is this MCP server free to use?", "acceptedAnswer": { "@type": "Answer", "text": "Yes. The free tier gives you 10 calls per day with no API key required. Pro tier is £79/mo for unlimited calls plus cryptographically signed attestations your auditor can verify independently." } }, { "@type": "Question", "name": "How does the signed attestation work?", "acceptedAnswer": { "@type": "Answer", "text": "Every Pro tier audit produces a HMAC-SHA256 signed certificate with a unique ID and a public verify URL. Your auditor pastes the cert into https://meok-attestation-api.vercel.app/verify and gets an independent valid/invalid response. No contact with MEOK required." } }, { "@type": "Question", "name": "Which MCP clients does this work with?", "acceptedAnswer": { "@type": "Answer", "text": "All standard MCP clients: Claude Desktop, Claude Code, Cursor, VS Code with MCP extension, Windsurf, Cline, and any custom MCP-compatible agent. Install via npx meok-setup or pip install for the underlying Python package." } }, { "@type": "Question", "name": "Can I install all MEOK governance MCPs at once?", "acceptedAnswer": { "@type": "Answer", "text": "Yes. Run npx meok-setup --pack governance to install all 10 governance MCPs and write the configs for Claude Desktop, Cursor, or Windsurf in one command." } }, { "@type": "Question", "name": "Is the regulation text authoritative?", "acceptedAnswer": { "@type": "Answer", "text": "Yes. MEOK syncs daily from the EUR-Lex Cellar SPARQL endpoint, the canonical EU regulation publication system. The text is verbatim with no LLM summarization. Every quote is auditor-defensible and includes the exact article number plus relevance score." } } ] } </script>