EU regulation database for AI agents

EU compliance, as MCP tools.

Tools agents call to check, cite, classify, and audit compliance against GDPR + AI Act + ePrivacy. Citation-backed answers from EUR-Lex source texts. Free tier for evaluation; paid tier for production.

Get an API key →Read the docs

See it in action — no signup

Pre-recorded real responses from our MCP server. Pick a tool, click an example.

Find articles + recitals by query — hybrid retrieval.

Pick an example above to run it.

outputs pre-generated 2026-05-19 · sign up to run your own queries

Why agents need this

Any agent or app touching EU users hits compliance questions — and the public regulation texts run to ~500 pages each. Pasting them into context burns tokens. ChatGPT-style answers sound right but cite nothing — or worse, cite articles that don't exist. Getting the cited article wrong can be career-ending for whoever ships the product.

euregs answers with citations grounded to the actual regulation text. Every reference is verified against the retrieved source — hallucinated articles never reach the output. Audit trail per call (paid tiers) so you can show regulators or auditors the chain of reasoning.

1.000

recall@3 on 24 seeded queries

586

articles + recitals indexed (GDPR · AI Act · ePrivacy)

MCP tools live — more on the roadmap

€29

Indie tier (5k calls/mo)

The toolbox

audit_document

Paste your privacy policy, ToS, cookie banner, or DPA — get a structured findings report. Missing clauses, weak language, suggested fixes — each grounded in a specific article.

paid

check_compliance

Describe an action, get a verdict (likely / uncertain / likely_not) + required actions, each grounded in specific articles. Hallucinated citations are filtered before you see them.

paid

classify_role

Describe your data flow → controller / processor / joint_controller. If the scenario is ambiguous, you get concrete questions about the missing facts — not a generic "uncertain".

paid

lookup_clause

Find the right article or recital across GDPR · AI Act · ePrivacy. Structural queries like article 9 or natural language like data minimization both work — top-3 recall is 100% on our eval set.

free + paid

Plus lookup_definition (resolve "data controller" / "GPAI" with aliases), resolve_references + find_references_to (citation graph in both directions), cite (Bluebook / OSCOLA / EU official formats), and enforcement_dates (phased applicability — AI Act high-risk applies 2026-08-02). Full docs →

Get started in 30 seconds

1. Sign up → email verification → API key shown once.

2. Add to Claude Desktop / Claude Code / Cursor:

claude mcp add euregs https://mcp.euregs.dev/mcp \
  --transport http \
  --header "Authorization: Bearer YOUR_KEY"

3. Ask:

> "Use euregs to check whether collecting EU user emails for marketing
   without explicit consent is GDPR-compliant. Cite specific articles."

Or call directly via HTTP — see the full docs for JSON-RPC payloads, rate limits, and audit trail behavior.

Pricing

Free tier: 100 calls/month, lookup + enforcement only. Sign up.

Indie €29/mo: 5,000 calls/month, all 8 tools including check_compliance, classify_role, audit_document. Pro €99/mo: 50,000 calls + premium synthesis quality. See full pricing →

Pay-per-call via x402 for autonomous agents — no account needed (coming soon).

Source + retention

Source corpus is EUR-Lex (the EU's official regulation publication — public domain). Every cited clause links back to the canonical text at eur-lex.europa.eu. Paid tiers include 12-month audit retention — every check_compliance call's input, retrieved chunks, raw LLM output, and validated output are stored for forensic evidence in case of compliance disputes.