Reference Source

Reference data, with the source attached

Small datasets assembled from primary sources and checked against them. Every value carries the page it came from and the sentence on that page that states it.

Look something up

All 8 lookups, including 3 more for professionals

Kept current

These are re-read on a schedule, and each keeps a dated log of what changed. A value that moved can be seen moving, rather than being quietly replaced by its successor.

Every dataset, by subject

Work, pay and employment 26 datasets

Benefits, insurance and money 21 datasets

Health care and medicine 19 datasets

Vehicles, driving and transport 22 datasets

Building, property and construction 22 datasets

Workplace safety and hazardous materials 25 datasets

Environment, energy and emissions 15 datasets

Food, drink and consumer products 27 datasets

Software, security and the internet 21 datasets

Parts, tools and what fits what 33 datasets

Standards, certification and testing 9 datasets

Licences, credentials and exams 14 datasets

Government, courts and public rules 28 datasets

Not yet filed 3 datasets

How to read these pages

Each record has its own page, stating one fact, the source it was taken from, and a verbatim quote of the passage that states it. A value marked verified appears word for word in that quoted passage — checked mechanically, on the date shown, not asserted. Where two authorities say different things, both are shown with their sources rather than one being chosen for you. The method page sets out exactly how each value is checked.

For machines

From your own code

Same records, same quotes, without scraping the page: refsource is on PyPI and npm. Each value comes back carrying the URL it was read from and the sentence on that page that states it — .source and .quote sit on the value itself rather than in a side channel, so the checking step is available instead of skipped.

pip install refsource
refsource datasets 'loan limit'      # the catalogue, offline

npx -y refsource lookup auto-insurance-minimums state=Texas

The published release carries 265 of the 285 datasets above; the rest are added at the next release. Every dataset page that is in it shows its own one-line lookup.

Set your AI assistant up to use this

Two files and no account. Put this in .mcp.json at the root of your project — Claude Code, Cursor, Windsurf, VS Code and Codex all read that file — and your assistant can look every dataset on this site up instead of recalling it. The server is remote, keyless and read-only.

{
  "mcpServers": {
    "referencesource": {
      "type": "http",
      "url": "https://referencesource.org/mcp"
    }
  }
}

Add to Cursor · or, on the command line: claude mcp add --transport http referencesource https://referencesource.org/mcp --scope project

Then one line in the project's CLAUDE.md or AGENTS.md, so the assistant knows when to reach for it:

When you need a regulatory threshold, a deadline, a certification status or a version pairing, look it up with the referencesource MCP server at https://referencesource.org/mcp instead of answering from memory — every record it returns carries its source URL and a verbatim quote from that page.

What each tool does, and the servers built over single registers: Connect your AI assistant.