Sovereign AI Act — Classificador MCP para EU AI Act
saidbazyar/sovereign-ai-act-mcp
Classificador EU AI Act determinístico para agentes — nunca adivinhas
Ver no GitHubMCP server que expõe Leo, um classificador determinístico para a Regulação (EU) 2024/1689, como ferramentas executáveis. Classifica sistemas de IA em categorias de risco (proibido, alto risco, limitado, mínimo) citando verbatim os Articles da lei oficial. Inclui tools para busca full-text, consulta de Articles específicos e prazos de conformidade.
Instalação
npx convez-ai saidbazyar/sovereign-ai-act-mcp1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Compliance e auditoria de sistemas de IA sob GDPR/EU AI Act
- Consulta automatizada de risco regulatório durante desenvolvimento
- Pesquisa e citação verbatim de Articles e Anexos da lei
Veredito

Ferramenta sólida e especializada para compliance regulatório na EU. Implementação transparente e determinística (não LLM-dependente) inspira confiança. Essencial para builders que precisam garantir conformidade com EU AI Act.
README
A tradução para o português deste README ainda está na fila.
◆ Sovereign AI Act — MCP server
Give your AI agent a trustworthy EU AI Act expert.
This Model Context Protocol server exposes Leo — the deterministic classifier for Regulation (EU) 2024/1689 (the EU AI Act) — as callable tools. Every answer is grounded verbatim in the official law. Leo never guesses.
When a user asks an AI "is my AI system high-risk under the EU AI Act?", the honest answer must cite the actual law — not a hallucination. This server lets your agent do exactly that.
🛠️ Tools
| Tool | What it does |
|---|---|
classify_ai_system | Plain-language AI description → risk tier (prohibited / high-risk / limited / minimal) + the exact Annex III category and binding Articles. |
lookup_article | Verbatim text of any Article (1–113). |
search_eu_ai_act | Full-text search across Articles, Recitals and Annexes. |
get_compliance_deadlines | The canonical post-Digital-Omnibus application dates and fine tiers. |
🚀 Install (Claude Desktop)
Add to your claude_desktop_config.json:
{
"mcpServers": {
"sovereign-ai-act": {
"command": "npx",
"args": ["-y", "sovereign-ai-act-mcp"]
}
}
}
Restart Claude, then ask: "Is an AI that screens job applicants high-risk under the EU AI Act?" — Claude calls Leo and answers with the exact Articles.
Works the same in Cursor, VS Code, Windsurf, Cline, and any MCP-capable client.
⚡ Try the API directly (no install, no key)
curl -X POST https://www.regulatoryai.eu/api/classify \
-H "content-type: application/json" \
-d '{"description":"AI that screens job applicants and ranks CVs"}'
{
"ok": true,
"tier": "high_risk",
"classifications": [{
"tier": "high_risk",
"rule": "Employment, recruitment & worker management (Annex III.4)",
"annex": "III.4",
"articles": [6, 9, 10, 11, 13, 14, 15],
"severity": "HIGH — conformity assessment + risk management + ..."
}]
}
Also: GET /api/article/{1-113} · GET /api/search?q=...
💎 Why trust it
- Verbatim law. All 113 Articles, 180 Recitals and 13 Annexes loaded word-for-word from the EU Publications Office (CELEX 32024R1689).
- Deterministic. A rule engine, not a guess — every verdict cites the Article that binds you.
- EU-sovereign. Operated by Dominion Intelligence AB, hosted in the EU.
- 24 languages. Ask in your own language.
⚙️ Config
SOVEREIGN_API_BASE— override the API base URL (defaulthttps://www.regulatoryai.eu).- No API key required.
Indicative classification grounded in the official text — not legal advice. Powered by Sovereign AI Act · © Dominion Intelligence AB