AI Act Radar Documentation
AI Act Radar delivers structured EU AI Act updates from 12 official sources via three interfaces: a webhook feed, a REST API, and an MCP server for AI agents.
Information only — not legal advice. AI Act Radar is a data and information service. It does not constitute legal advice and is not a substitute for qualified legal counsel.
Where to start
- Quickstart — make your first authenticated request in under 60 seconds.
- Authentication — Bearer tokens, rotation, and rate limits.
- Webhook payload schema — what lands at your endpoint.
- Signature verification — HMAC-SHA256, in your language of choice.
- MCP server — drop-in tools for Cursor, Claude Desktop, agent stacks.
Three layers, one source of truth
| Layer | Use when | Example |
|---|---|---|
| Webhook | You want push: real-time delivery to your endpoint when a topic-matching update lands. | payload schema |
| REST API | You want pull: query historical events, run filters, build a dashboard. | GET /v1/events |
| MCP server | You operate an AI agent and want it to ask the AI Act questions directly. | query_ai_act_updates |
What's monitored
12 primary sources, fan-in to one normalised schema:
- Primary law: EUR-Lex (CELEX 32024R1689 family), Official Journal
- EU bodies: AI Office, AI Board, GPAI Code of Practice, Have-Your-Say
- Standards: CEN/CENELEC JTC 21
- National regulators: BNetzA, BfDI, BSI (DE), CNIL (FR), Algoritmeregister (NL)
- Cross-country: OECD.AI Policy Observatory
See sources & cadence for fetch intervals and licence info.
Disclaimers everywhere
Every webhook payload, every API response, and every MCP tool call carries
a disclaimer field stating that the content is informational
and not legal advice. We also set
X-AIAR-Disclaimer: informational-only-not-legal-advice as an
HTTP header on every response. If you forward our events to humans, please
carry the disclaimer along.