{
  "markdown": "# Czech Law MCP\n\n<!-- ANSVAR-CTA-BEGIN -->\n> **The Czech law corpus is now served through the Ansvar Gateway.** Connect your AI assistant (Claude, Copilot, Cursor, custom MCP client) to `https://gateway.ansvar.eu/mcp` — one OAuth connection, free tier available, covering this corpus plus EU regulations, national law across 28 audited jurisdictions, and CVE/security intelligence, every result with a verbatim source citation. Start at https://ansvar.eu/docs/quickstart\n\n### Connect\n\n**Claude Code** (one line):\n\n```bash\nclaude mcp add ansvar --transport http https://gateway.ansvar.eu/mcp\n```\n\n**Claude Desktop / Cursor** — add to `claude_desktop_config.json` (or `mcp.json`):\n\n```json\n{\n  \"mcpServers\": {\n    \"ansvar\": {\n      \"type\": \"url\",\n      \"url\": \"https://gateway.ansvar.eu/mcp\"\n    }\n  }\n}\n```\n\n**Claude.ai** — Settings → Connectors → Add custom connector → paste `https://gateway.ansvar.eu/mcp`\n\nFirst request opens an OAuth signup flow (setup details: [ansvar.eu/docs/quickstart](https://ansvar.eu/docs/quickstart)). After signup, your client is bound to your account; tier (free / premium / team / company) determines fan-out, quota, and which downstream MCPs are reachable.\n\n---\n\n## Self-host this MCP\n\nYou can also clone this repo and build the corpus yourself. The schema,\nfetcher, and tool implementations all live here. What is not in the repo is\nthe pre-built database — TDM and standards-licensing constraints on the\nupstream sources mean we host the corpus on Ansvar infrastructure rather\nthan redistribute it as a public artifact.\n\nBuild your own: run this repo's ingestion script (entry-point varies per\nrepo — typically `scripts/ingest.sh`, `npm run ingest`, or `make ingest`;\ncheck the repo root).\n<!-- ANSVAR-CTA-END -->\n\n\nMCP server for Czech Law — 45,899 statutes from www.e-sbirka.cz.\n\n[![License: Apache 2.0](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](LICENSE)\n[![MCP](https://img.shields.io/badge/MCP-spec--compliant-green.svg)](https://modelcontextprotocol.io)\n[![Jurisdiction](https://img.shields.io/badge/Jurisdiction-CZ-informational.svg)](#coverage)\n\n## What this is\n\nThis server indexes the legal materials listed under **Sources** below and\nexposes them via the Model Context Protocol. Part of the Ansvar MCP fleet —\nsource-available servers published for self-hosting.\n\nIt makes no outbound network calls except to the upstream sources during\ningestion — no analytics, no phone-home.\n\n## Coverage\n\n- **Corpus:** Czech Law — 45,899 statutes, 461,231 provisions\n- **Jurisdiction code:** `CZ`\n- **Corpus snapshot:** 2026-02-22\n\nThe corpus is rebuilt from the upstream sources by the included ingestion script; re-run periodically to refresh.\n\nSee **Sources** below for source URLs, terms, and reuse conditions.\n\n## Why this exists\n\nLLMs answering compliance, security, or legal questions from training data\nalone will fabricate citations — confidently producing article numbers,\nstatute names, and source URLs that do not exist, or that do not say what\nthe model claims. This MCP exists so an agent can call a tool that returns\nthe real text, the real identifier, and the real source URL straight from\nthe indexed materials — and ground an answer rather than recall it.\n\nOne MCP, one corpus. The point is composition.\n\nThe **Ansvar Gateway** ([ansvar.eu](https://ansvar.eu)) joins this MCP\nwith the rest of the Ansvar fleet behind a single authenticated\nendpoint — 300+ servers covering legal jurisdictions, EU\nregulations, security frameworks, sector regulators, privacy-pattern\ncatalogues, and risk-scoring tools. That lets an agent run cross-domain\nworkflows that no single MCP can serve alone:\n\n- **Threat model and TARA.** Threat enumeration → known component\n  vulnerabilities → severity scoring → applicable AI, cybersecurity, and\n  automotive obligations → privacy threats. Every finding traceable to\n  its source.\n- **Gap analysis.** Target framework requirements → current-state\n  evidence → unmet obligations → remediation guidance and authority\n  opinions. Every gap traceable to the specific requirement that flagged\n  it.\n- **Data Protection Impact Assessment.** Privacy regulation articles →\n  national DPA guidance → privacy-pattern catalogue → applicable case\n  law.\n\n### Getting high-quality citations\n\nCitation accuracy degrades when an agent's context fills up. Long inputs\ncause retrieval-stage drift — the model recalls claim text correctly but\nmisattributes the source. Two practices keep accuracy high:\n\n1. **Focused first pass, checking-agent second pass.** Query a small,\n   relevant set of MCPs first, then run a separate agent whose only job\n   is to re-resolve each citation against the source MCP and flag any\n   that no longer match. The checking agent uses the same MCP tools as\n   the synthesis agent.\n2. **Pull the source text verbatim when in doubt.** Every citation an\n   agent emits points back to a tool call against this server. You — or\n   another agent — can call the same tool with the same identifier and\n   read the raw statute, article, or standard text directly. If the\n   verbatim text doesn't support what the agent claimed, the citation\n   was misused, regardless of whether the identifier was real.\n\nBoth patterns work the same way self-hosted or through the gateway.\n\n## Two ways to use it\n\n**Self-host (free, Apache 2.0)** — clone this repo, run the ingestion script to build your local database from the listed upstream sources, point your MCP client at the local server. Instructions below.\n\n**Use the hosted gateway** — for production use against the curated,\nkept-fresh corpus across the full Ansvar MCP fleet, with citation enrichment\nand multi-jurisdiction fan-out — see [ansvar.eu](https://ansvar.eu).\n\n## Self-hosting\n\n### Install\n\n```bash\ngit clone https://github.com/Ansvar-Systems/Czech-law-mcp.git\ncd Czech-law-mcp\nnpm install\n```\n\n### Build the database\n\n```bash\nnpm run build:db\n```\n\nIngestion is a snapshot — your local copy goes stale until you re-run it. The hosted gateway corpus is refreshed continuously.\n\n### Premium tier (case-law)\n\nThe premium tier extends the free corpus with curated Supreme Court decisions from the **Sbírka soudních rozhodnutí a stanovisek** published at [sbirka.nsoud.cz](https://sbirka.nsoud.cz). License basis is the same Czech-Statutory-PD §3 carve-out that covers statutes — court decisions are excluded from copyright protection under Czech Copyright Act 121/2000 Coll.\n\nBuild the premium database in two steps. First, ingest the case-law corpus (single-threaded, ≥ 2 seconds per request, honouring `robots.txt`):\n\n```bash\nnpm run ingest:nsoud-sbirka                  # full corpus (~9,790 decisions, ~6 hours)\nnpm run ingest:nsoud-sbirka -- --limit 100   # smoke test (200 seconds)\nnpm run ingest:nsoud-sbirka -- --resume      # skip already-ingested URLs\n```\n\nSeed files land under `data/case-law-seed/{ns-sbirka-<id>}.json`. Then build the premium database, which combines the e-Sbírka statute corpus with the case-law seeds:\n\n```bash\nnpm run build:db:premium\n```\n\nThe premium build writes to `data/database-premium.db` (~1.3 GB statutes + case-law growth). Case-law rows are stored in the same `legal_documents` table with `type='case_law'`, `publisher='Nejvyšší soud'`, `license='Czech-Statutory-PD'`. Cited statutes parsed from each decision's `Předpisy` field land in `cross_references`.\n\nThe full Supreme Court decision search at `rozhodnuti.nsoud.cz` is out of scope — that subdomain blocks all non-DG_JUSTICE_CRAWLER user-agents via `robots.txt` and requires a separate written exception from NS IT. The Sbírka collection is the curated, publication-grade subset.\n\n### Configure your MCP client\n\n```json\n{\n  \"mcpServers\": {\n    \"czech-law-mcp\": {\n      \"command\": \"node\",\n      \"args\": [\"dist/index.js\"]\n    }\n  }\n}\n```\n\n## Sources\n\n| Source | Source URL | Terms / license URL | License basis | Attribution required | Commercial use | Redistribution / caching | Notes |\n|---|---|---|---|---|---|---|---|\n| [www.e-sbirka.cz](https://www.e-sbirka.cz) | https://www.e-sbirka.cz | [Czech Copyright Act 121/2000 §3](https://www.e-sbirka.cz/eli/cz/sb/2000/121/) | `Czech-Statutory-PD` — Czech Copyright Act 121/2000 Coll. §3 excludes \"úřední dílo\" (official works) from copyright protection | Yes | Yes | Yes | Free tier. Same statutory carve-out mechanism as US-Federal-PD (17 USC §105) and Norwegian-Court-Publication. Czech transposition of EU Open Data Directive 2019/1024 (Act 261/2021 Coll.) provides supplementary basis. Catalogued upstream as `Czech-Statutory-PD` (entry_kind: regime). |\n| [sbirka.nsoud.cz](https://sbirka.nsoud.cz) | https://sbirka.nsoud.cz/sbirka/{id}/ | [Czech Copyright Act 121/2000 §3](https://www.e-sbirka.cz/eli/cz/sb/2000/121/) | `Czech-Statutory-PD` — same úřední dílo carve-out applied to court decisions | Yes | Yes | Yes | Premium tier (Wave A). Curated Supreme Court decisions (Sbírka soudních rozhodnutí a stanovisek). `robots.txt` allows full crawl with no Crawl-Delay; ingestion is single-threaded at ≥ 2 s per request. Full Supreme Court search at `rozhodnuti.nsoud.cz` blocks non-DG_JUSTICE_CRAWLER agents and is out of scope. |\n\n## What this repository does not provide\n\nThis repository's source — the MCP server code, schema, and ingestion script — is licensed under Apache\n2.0. The license below covers the code in this repository only; it does not\nextend to the upstream legal materials.\n\nRunning ingestion may download, cache, transform, and index materials from the listed upstream sources. You are responsible for confirming that your use of those materials complies with the source terms, attribution requirements, robots/rate limits, database rights, copyright rules, and any commercial-use or redistribution limits that apply in your jurisdiction.\n\n## License\n\nApache 2.0 — see [LICENSE](LICENSE). Commercial use, modification, and\nredistribution of **the source code in this repository** are permitted under\nthat license. The license does not extend to upstream legal materials downloaded by the ingestion script; those remain governed by the source jurisdictions' own publishing terms (see Sources above).\n\n## The Ansvar gateway\n\nIf you'd rather not self-host, [ansvar.eu](https://ansvar.eu) provides this\nMCP plus the full Ansvar fleet through a single authenticated endpoint, with\nthe curated production corpus, multi-MCP query orchestration, and citation\nenrichment.\n\n---\n\nIssues: [github.com/Ansvar-Systems/Czech-law-mcp/issues](https://github.com/Ansvar-Systems/Czech-law-mcp/issues) · Security: <security@ansvar.eu>\n",
  "bytes": 10537,
  "sha": "98d9999c7a6746eb7081cdcf412fa13e9f95c60782f85d70bef3689a75eb2a41",
  "repo_slug": "ansvar-systems/czech-law-mcp",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_eu_ansvar_czech_law_mcp_0169f360/readme"
}