German Legal MCP Server
Provides unified access to German federal and state legislation, court decisions, and European Union legal databases for comprehensive legal research. It enables users to search and retrieve full-text laws, parliamentary documents, and judicial rulings directly through the Model Context Protocol.
README
German Legal MCP Server
⚠️ WARNING: Work in Progress
This project is currently under active development and not production-ready. APIs may change without notice, and features may be incomplete or unstable. Use at your own risk.
A Model Context Protocol (MCP) server for German legal research, providing unified access to multiple legal databases.
Supported Sources
| Source | Status | Prefix | Authentication |
|---|---|---|---|
| Bundes- & Landesrecht | ✅ Available | legis: |
None (public) |
| Rechtsprechung im Internet | ✅ Available | rii: |
None (public) |
| InfoCuria (CJEU) | ✅ Available | icu: |
None (public) |
| EUR-Lex | ✅ Available | eul: |
None (public) |
| DIP Bundestag | ✅ Available | dip: |
Public key included |
| arXiv | ✅ Available | arxiv: |
None (public) |
Features
Bundes- & Landesrecht (legis:* tools)
- Federal and state legislation — BUND (all federal laws) + 16 Länder (all states)
- No authentication — free public access, no rate limits
- Unified interface — one set of tools for all jurisdictions
- Full text search — search across state legislation (Länder only)
- Resilient input — BUND accepts "§ 823", "823", "Art. 1", "Paragraph 51"
- Pandoc-compatible Markdown — clean conversion with Turndown
- Save to file —
save_pathparameter to avoid context pollution - Available states: BUND, BB, BW, BY, BE, HB, HE, HH, MV, NI, NW, RP, SL, SN, ST, SH, TH
Rechtsprechung im Internet (rii:* tools)
- Federal court decisions — BVerfG, BGH, BVerwG, BFH, BAG, BSG, BPatG (from 2010)
- Bavarian state courts — AG, LG, OLG, VG, VGH, FG, ArbG, LAG, BayVerfGH via gesetze-bayern.de
- No authentication — free public access
- Full text search — search across all federal court decisions
- Kurztext/Langtext — summary or full text via
partparameter - Randnummern — formatted as
[Rn. 5]{.rn}(pandoc spans) - Save to file —
save_pathparameter to avoid context pollution
InfoCuria — CJEU (icu:* tools)
- EU Court of Justice case law — judgments, opinions, orders from CJEU and General Court
- No authentication — free public access via InfoCuria API
- Multilingual — documents available in all EU languages (default: DE)
- Flexible case lookup — accepts case numbers (C-476/17), CELEX numbers, or internal IDs
- Randnummern — formatted as
[Rn. 5]{.rn} - Partial content —
sectionparameter for Rn ranges, headings, or line ranges - Save to file —
save_pathparameter to avoid context pollution
EUR-Lex (eul:* tools)
- EU legislation — directives, regulations, decisions, treaties (TFEU, TEU)
- No authentication — free public access via Cellar REST API and SPARQL
- Multilingual — documents available in all EU languages (default: DE)
- CELEX lookup — retrieve by CELEX number (e.g., "32016R0679" for GDPR)
- SPARQL search — search by title keywords, filter by resource type
- Partial content —
sectionparameter for articles (Art. 5), headings, or line ranges - Save to file —
save_pathparameter to avoid context pollution
DIP Bundestag (dip:* tools)
- Parliamentary documents — Bundestagsdrucksachen (Gesetzentwürfe, Beschlussempfehlungen, Anfragen)
- Legislative processes — Vorgänge with status tracking and linked documents
- Debate transcripts — full text search across Plenarprotokolle (BT and BR)
- Full text retrieval — extracted text including Gesetzesbegründungen, with section support
- Public API key included — works out of the box (key expires 2026-06-01, override via env var)
- Save to file —
save_pathparameter to avoid context pollution
arXiv (arxiv:* tools)
- Preprint search — search by keywords, author, title, abstract, or category
- Metadata + abstract — default response without full text fetch (token-efficient)
- HTML full text — Markdown conversion for papers from ~2024+ (LaTeXML HTML)
- PDF fallback — older papers without HTML return abstract + PDF link
- No authentication — free public API, no rate limits beyond ~1 req/3s
- Save to file —
save_pathparameter to avoid context pollution
Quick Start with npx
npx @metaneutrons/german-legal-mcp
or add your MCP client config (e.g., claude_desktop_config.json):
{
"mcpServers": {
"german-legal": {
"command": "npx",
"args": ["-y", "@metaneutrons/german-legal-mcp"]
}
}
}
Environment Variables
Provider Enablement
| Variable | Default | Description |
|---|---|---|
GLMCP_LEGIS_ENABLED |
true |
Bundes- & Landesrecht |
GLMCP_RII_ENABLED |
true |
Rechtsprechung im Internet |
GLMCP_ICU_ENABLED |
true |
InfoCuria (CJEU) |
GLMCP_EUL_ENABLED |
true |
EUR-Lex |
GLMCP_DIP_ENABLED |
true |
DIP Bundestag (auto-disabled after 2026-06-01 without own key) |
GLMCP_DIP_API_KEY |
Public key | Override the bundled public API key |
GLMCP_ARXIV_ENABLED |
true |
arXiv preprint search |
Tools
Bundes- & Landesrecht
| Tool | Description |
|---|---|
legis:search |
Search federal and state legislation. Parameter: query, state (e.g., "BW", "BE"), limit. Note: BUND does not support search. |
legis:get |
Retrieve a specific law/norm. BUND: id = "law/section" (e.g., "bgb/823"). Länder: id from search results. Optional save_path. |
legis:toc |
Compact table of contents for a law — section numbers and headings. Supports from/to range and depth filter. BUND: id = law abbreviation (e.g., "bgb"). |
legis:states |
List available jurisdictions with implementation status. |
Rechtsprechung im Internet
| Tool | Description |
|---|---|
rii:search |
Search for court decisions. Returns list with doc IDs, titles, and snippets. Use source: "BY" for Bavarian state courts. |
rii:get_decision |
Retrieve full text of a court decision by doc ID. part: K (Kurztext) or L (Langtext, default). Optional save_path to save to file. Use source: "BY" for gesetze-bayern.de IDs. |
InfoCuria — CJEU
| Tool | Description |
|---|---|
icu:search |
Search CJEU decisions and opinions. Returns case numbers, ECLI, dates, and document IDs. |
icu:get_document |
Retrieve full text by case number (C-476/17) or CELEX number. Supports section (Rn ranges, headings, line ranges) and save_path. |
EUR-Lex
| Tool | Description |
|---|---|
eul:search |
Search EU legislation via SPARQL. Filter by type (directive, regulation, decision, treaty). |
eul:get_document |
Retrieve EU legislation by CELEX number (e.g., "32016R0679" for GDPR). Supports section (Art. 5, Artikel 5-10, headings, line ranges) and save_path. |
DIP Bundestag
| Tool | Description |
|---|---|
dip:search |
Search Bundestagsdrucksachen by title. Filter by type (Gesetzentwurf, Anfrage, etc.), Wahlperiode, date range. |
dip:get |
Retrieve full text of a Drucksache by Dokumentnummer (e.g., "19/27426"). Supports section and save_path. |
dip:search_vorgang |
Search legislative processes (Vorgänge) with status and linked Drucksachen. |
dip:search_plenarprotokoll |
Full text search across parliamentary debate transcripts (BT and BR). |
arXiv
| Tool | Description |
|---|---|
arxiv:search |
Search preprints by keywords, author, title, abstract, or category. Returns metadata + abstract. |
arxiv:get |
Retrieve paper by arXiv ID. Default: metadata + abstract. With section or save_path: HTML full text as Markdown (~2024+, older: PDF link). |
Two-Phase Document Retrieval
All document tools use a two-phase approach to avoid flooding the LLM context:
- Outline — first call returns title, metadata, table of contents, and a preview
- Section — request specific parts by Randnummer, heading, or line range (served from cache)
- Save to file — write full document to disk, return metadata only
Section formats: "Rn 5", "Rn 5-12", "lines:100-200", or any heading text (fuzzy match).
Markdown Output
Documents are converted to pandoc-compatible Markdown:
- Randnummern:
[Rn. 5]{.rn}(bracketed spans) - Footnotes:
[^1]references with[^1]: textdefinitions
Development
npm test # Run tests
npm run test:watch # Watch mode
npm run test:coverage # Coverage report
MCP Inspector
npx @modelcontextprotocol/inspector node dist/index.js
Commit Convention
This repo uses Conventional Commits enforced via Husky + commitlint.
Types: feat, fix, docs, style, refactor, perf, test, chore, ci, build, revert
Scopes: legis, rii, icu, eul, dip, core, deps, config
Architecture
- Dynamic provider loading — providers auto-discovered from
src/providers/*/ - Cheerio + Turndown for HTML → pandoc Markdown conversion
- Zod for input validation
- Axios for HTTP requests (Legis, RII, InfoCuria, EUR-Lex, DIP, arXiv)
- Tools namespaced by source (
legis:,rii:,icu:,eul:,dip:,arxiv:)
License
GPL-3.0 - See LICENSE for details.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.