Connect your AI assistant to AI regulation data

A free MCP server giving Claude, ChatGPT or any other assistant live access to 1,142 AI laws across 110 countries and jurisdictions, 280 enforcement actions, and a 4,241-term glossary. No account, no API key.

https://regulations.ai/api/mcp/v1

What it answers

A general-purpose model knows AI law up to its training cut-off and cannot tell you what changed last week, or which official source a status came from. This server can, because every answer is drawn from a corpus that is rescanned continuously and carries the official source for each entry.

  • →Which AI laws apply in a given country, and what stage each one is at
  • →What has changed in the last 30 days
  • →How different laws define the same term — "high-risk AI system" is defined 98 separate ways
  • →Which regulators have taken enforcement action, and against whom

How to connect

Claude

Desktop or claude.ai
  1. Open Settings (Ctrl+, on desktop, or your profile icon on the web)
  2. Click Connectors in the sidebar
  3. Click Add, then Add custom connector
  4. Paste https://regulations.ai/api/mcp/v1 and click Add

There is no authentication step — skip straight past it. Note that this will NOT work from claude_desktop_config.json: that file starts local programs, and cannot take a remote URL.

ChatGPT

Developer mode
  1. Open Settings, then Security and login
  2. Turn on Developer mode
  3. Go to Plugins and click the + button
  4. Enter https://regulations.ai/api/mcp/v1

This server implements the search and fetch tools ChatGPT requires of a connector, so results come back as citations linking to the underlying record.

Any other MCP client

Cursor, Zed, Claude Code, your own agent
  1. Add a remote MCP server in your client's configuration
  2. Transport: streamable-http
  3. URL: https://regulations.ai/api/mcp/v1
  4. No credentials required

The server speaks JSON-RPC 2.0 over Streamable HTTP and supports protocol versions 2025-06-18, 2025-03-26 and 2024-11-05.

Tools

search
Search everything — laws, definitions, enforcement, research, news. Returns ids, titles and links.
fetch
Retrieve one record by id: the key facts, a summary, and the page it came from.
search_regulations
Filter AI laws by jurisdiction, status or instrument type.
get_regulation
One instrument: jurisdiction, status, dates, official source.
recent_changes
What has moved recently. A general model cannot answer this — the corpus is rescanned weekly.
get_glossary_term
A term of art, with how many separate laws define it and how they differ.
list_jurisdictions
Every jurisdiction covered, with counts.
compare_regulations
Two to five instruments side by side.

Limits and licence

Without a key, results carry the facts about an instrument — jurisdiction, status, type, dates, official source — plus a short extract and a link to the full record. Up to 5 results per search, and a 90-day window on recent changes. That is deliberate: the complete analysis, the sources and the compliance timeline live on the page, and this server is meant to point people there rather than replace it.

A free API key raises the limits to 50 results and a 365-day window, and returns the full summary text.

Please cite the record URL when you use an answer. Every response includes one, and it is the only way a reader can check a claim against the official source behind it.

Also available

  • /llms.txt — a plain-text map of the site for AI crawlers
  • /okf/index.md — the corpus as an Open Knowledge Format bundle, one file per jurisdiction
  • /enterprise — REST endpoints for partners who would rather not speak MCP