eu-law-mcp
# eu-law-mcp
Read-only MCP server for exact retrieval of authoritative EU legislation, case law, and regulator material. Results include official identifiers, provenance, durable evidence, and source anchors.
[Documentation](https://hatt-io.github.io/eurmcp/)
## Install
Requires Node.js 20+ and npm 11.18+.
```bash
git clone https://github.com/hatt-io/eurmcp.git eu-law-mcp
cd eu-law-mcp
npm ci
npm run build
```
## Launch
From the cloned repository root, configure a stdio MCP client to run:
```bash
node dist/index.js
```
The server provides 16 read-only tools for:
- legislation search and exact document, article, and recital retrieval;
- official versions, timelines, point-in-time provisions, and structural diffs;
- case search, exact judgments and paragraphs, and citation relationships;
- EDPB and EDPS material;
- anchored quote verification against stored official evidence.
See the [MCP tool reference](https://hatt-io.github.io/eurmcp/02%20Reference/MCP%20Tools.html) for inputs, outputs, and client setup.
## Verify
```bash
npm run verify
```
Live official-source tests are opt-in:
```bash
npm run test:live
```
## License
MIT
TDQS
Scored across 16 tools
Every tool targets a distinct retrieval or search operation with clear domain boundaries: legislation vs. case law vs. EDPB vs. EDPS. Even similar operations like get_article and get_provision_at_date are differentiated by identifier vs. date, and version-related tools (timeline, list, compare) have non-overlapping purposes.
All 16 tools follow a consistent verb_noun snake_case pattern (get_, search_, compare_, find_, list_, verify_). The verbs are descriptive and uniform, making the tool surface predictable and easy to navigate for an agent.
With 16 tools, the server is slightly above the optimal 3–15 range but still well-scoped for its multi-domain purpose (legislation, case law, EDPB, EDPS). Each tool earns its place, covering distinct operations without unnecessary bloat.
The set provides comprehensive coverage for a read-only legal information server: full-document retrieval, granular provision access (articles, recitals, paragraphs), version history and comparison, date-based access, and citation verification. No obvious dead ends or missing core operations are apparent.