Skip to main content
Glama

legalize-mcp

One MCP server, 37 jurisdictions. A read-only Model Context Protocol server over the legalize-dev corpus — national legislation stored as law-as-git: one law per Markdown file, every reform a Git commit, each with an 8-field ELI-style frontmatter (Legalize Format Spec v0.2). MIT-licensed source corpus; this connector is Apache-2.0.

Built by Matematic Solutions as part of the eu-legal-mcp line. Unlike the single-country *-eli-mcp connectors, this one serves many jurisdictions from a single server, because the underlying corpus is already normalised to one format.

Why this exists

Every response carries the citation contract the rest of the line uses:

  • source_url — the official government source from the law's own frontmatter (e.g. boe.es, riigiteataja.ee, legifrance.gouv.fr). Never invented.

  • github_url — the verifiable legalize-dev copy actually read.

  • human_readable_citation"<title> (<identifier>)".

It is RODO/GDPR-safe and self-hostable: nothing leaves the user's machine except read-only GETs to GitHub.

Related MCP server: Vaquill-AI/vaquill-mcp

Jurisdictions

EU: at be cz de ee es eu fi fr gr ie it lt lu lv nl pl pt ro se sk · Other: ad ar ch cl co kr li no uk us uy

Call legalize_list_countries for the live list (add eu_only=true to filter).

Tools

Tool

Keyless?

Purpose

legalize_list_countries

List jurisdictions (code, name, EU flag, repo).

legalize_search_laws

needs GITHUB_TOKEN

Keyword search inside a country's laws → law_id + snippet.

legalize_get_meta

Frontmatter only (cheap citation check).

legalize_get_law

Full metadata + text by law_id; pass sha= for a historical version.

legalize_list_reforms

Reform timeline (commits) for a law; SHAs feed legalize_get_law.

law_id is the Markdown filename stem (e.g. BOE-A-1978-31229 for the Spanish Constitution, a numeric id for Estonia). Get it from legalize_search_laws or an official citation — there is no fuzzy title lookup.

Install

uvx legalize-mcp           # run directly
# or
pip install legalize-mcp

Claude Code (.mcp.json):

{
  "mcpServers": {
    "legalize": { "command": "uvx", "args": ["legalize-mcp"] }
  }
}

GITHUB_TOKEN (or GH_TOKEN) is optional — required only for legalize_search_laws, and it lifts the GitHub rate limit on legalize_list_reforms.

Configuration

Env var

Default

Purpose

GITHUB_TOKEN / GH_TOKEN

Enables code search; lifts rate limits.

LEGALIZE_AUDIT_DIR

~/.matematic/audit

JSONL audit log location (AI Act art. 12).

Governance

Every tool call appends one line to ~/.matematic/audit/legalize-mcp.jsonl (input hash, duration, status — no payloads). All tools are read-only and idempotent. See CONSTITUTION.md.

Development

python -m venv .venv && .venv/Scripts/python -m pip install -e ".[dev]"
.venv/Scripts/python -m pytest -q          # offline unit tests
.venv/Scripts/python -m ruff check src tests

Attribution

Legislative data © the respective national authorities (see each law's source). Corpus normalisation by the legalize-dev project under the MIT License. This connector does not redistribute the corpus; it reads it on demand.

A
license - permissive license
-
quality - not tested
C
maintenance

Maintenance

Maintainers
Response time
Release cycle
Releases (12mo)
Commit activity

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/matematicsolutions/legalize-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server