GitHub radar

Check any country's law on any date — stop your AI agent guessing

This is a repository with the dated history of laws from many countries: your AI agent pulls exact legal text from this database instead of its memory, and tells you exactly when it took effect.

01lexifina/global-law 1.5kHTML

A git repository (a tool that tracks file history) holding versions of laws from different countries, where each change is a dated entry for when it took effect. Alongside it sits a local server that lets an AI agent read this database instead of relying on memory.

When it helps

Useful if you're building an agent app or just want to double-check a legal quote for a specific date. Not a lawyer substitute — it's a way to verify citations, not legal advice.

Pros

  • Exact in-force dates for laws, not the model's guesses.
  • 18 ready-made tools to search and compare laws.
  • Runs locally and makes no network calls.
  • Free CC0 license — use it without restrictions.

Cons

  • Many countries aren't fully covered — some have no law text.
  • Some laws show only the current version, with no history.
  • Not legal advice — only a tool for checking citations.

How to set it up — step by step

  1. 1Install Python 3.10+ and git if you don't have them yet.
  2. 2Clone the repo with `git clone --recurse-submodules https://github.com/lexifina/global-law`.
  3. 3Open Claude Code in the repo folder and run `claude mcp add global-law -- python tools/mcp_server.py`.
  4. 4Send the agent the text from the block below.
  5. 5Check that the agent cites a specific date and law, not just memory.

Text for your agent

Copy this and send it to your agent — Claude Code, Codex, any of them:

Connect the global-law MCP server from https://github.com/lexifina/global-law — clone it with `git clone --recurse-submodules`, then run `claude mcp add global-law -- python tools/mcp_server.py` in Claude Code, and ask it what a specific legal provision (say, in your own country) said on a given date. Show me what it answers.

Runs on a regular laptop — just Python 3.10+ and git, no GPU needed; the server runs locally only and makes no network calls.

Open on GitHub