법제처 국가법령정보 MCP — 법령·판례·조례 조회부터 인용 환각 검증까지 · Korean law MCP for LLMs
2.2k
Stars
413
Forks
1
Open issues
6
Contributors
AI Analysis
Korean Law MCP is a Model Context Protocol server that integrates South Korea's National Law Information System (법제처) APIs to provide LLMs with tools for querying laws, precedents, ordinances, and treaties, with built-in hallucination detection via citation verification and content matching. It serves legal AI assistants, compliance specialists, and government officials who need reliable Korean legal research integrated into Claude, Cursor, or other AI platforms—not suitable for general progr...
Inferred from signals mentioned in the README (tests, CI, type safety) — not a review of the actual code.
AI's overall editorial judgment — not an average of the bars above, can weigh other factors too.
Korean legal research MCP wrapping 42 government APIs into 9 structured tools for citation verification and statutory analysis
korean-law-mcp is a Model Context Protocol server that exposes South Korea's National Law Information database (법제처) through Claude, Cursor, and other MCP-capable clients. It wraps 42 Korean government legal APIs into 9 consolidated tools focusing on preventing LLM hallucination in legal citations, tracing precedent validity (citator function), determining applicable historical statute versions, and mapping regulatory cross-references. Adoption appears concentrated in Korean legal technology, compliance, and research communities.
Project created 2025-12-19, representing the recent wave of MCP-based integrations of specialized data sources. Appears purpose-built to address a specific friction: Korean legal AI applications requiring real-time verification against authoritative government sources to prevent fabricated statutes and outdated law citations in high-stakes contexts.
99 stars in 7 days (snapshot at analysis date 2026-07-01) indicates rapid initial traction, likely driven by Korean developer communities and legal tech circles. 2,086 stars and 393 forks in ~6.5 months suggest adoption as a reference implementation for Korean legal API integration and possibly as a production tool in localized AI assistants. Growth pattern consistent with a specialized-domain tool gaining foothold among practitioners who recognize its specific value, not a general-audience project.
adoption not verified. No explicit case studies, customer testimonials, or documented production deployments in README. Top-level badge shows '조코헌트 주간 1등' (weekly winner on Korean tech community platform), suggesting community recognition but not confirmed commercial/institutional usage. GitHub star velocity and fork count are signals of interest, not proof of production deployment.
Based on README: TypeScript MCP server exposing procedural tools organized by legal research task (legal_research with 8 task modes: full_research, law_system, action_basis, dispute_prep, amendment_track, ordinance_compare, procedure_detail, document_review) and legal analysis modes (legal_analysis with cite_check, applicable_law, impact_map, verify_citations). V4.4 consolidated 19 tools to 9 to reduce tool-listing context overhead (52% reduction). Appears to batch-call government APIs and parse XML/JSON responses, with Zod schemas for validation. Integrates fallback sources (국세청 tax service for precedent texts when primary source lacks body text).
not documented in README
Last push 2026-06-30 (1 day before analysis date), indicating active maintenance. README shows rapid iteration: v4.0 through v4.4 span ~5 months with specific bug fixes (Zod version pinning, schema `required` field bugs, XML parsing edge cases, API Referer header injection, retry logic for intermittent blank/HTML responses). Multiple security patches (v4.0.5: @xmldom/xmldom injection, @hono IPv6 bypass, fast-uri traversal). Pattern suggests responsive development to real operational friction, not passive maintenance.
ADOPT IF: you are building Korean legal AI applications, compliance automation, or research systems where citation accuracy and statutory historical versions are non-negotiable; your primary users speak Korean and operate within South Korean regulatory frameworks; you have access to an MCP-capable client (Claude, Cursor, Windsurf, Zed, or Claude.ai). AVOID IF: you need commercial SLA/support, curated legal analysis beyond API data, or multi-jurisdictional coverage; you require precedent citator features matching Westlaw/LexisNexis editorial standards; you are a non-Korean legal service (API is Korean-language and Korean law specific). MONITOR IF: you are evaluating whether to build Korean legal tooling in-house vs. adopting this; recent patch velocity suggests maintainer is responsive to operational bugs, but long-term maintenance beyond current developer is uncertain.
Independent dimensions
Mainstream potential
3/10
Technical importance
6/10
Adoption evidence
2/10
- Dependency on 法制處 (Korean government) API availability and schema stability — no SLA documented; v4.0.8 already encountered intermittent blank/HTML responses requiring retry logic.
- Single-maintainer appearance (chrisryugj GitHub); sustainability of support and feature development unclear if primary developer moves to other priorities.
- Adoption not verified at commercial scale — stars and forks may represent experimentation rather than production deployments; no documented rollout at legal firms, government agencies, or enterprise AI platforms.
- Korean-language-only documentation and README; limits discoverability and contribution from non-Korean developer communities, may create support bottleneck.
- LLM hallucination prevention features (cite_check, applicable_law, impact_map) assume government API data is ground truth; tool's value is contingent on confidence in official sources, which may shift with regulatory changes or data quality issues.
Likely to remain a specialized, well-maintained tool for Korean-language legal AI workflows. If Korean legal tech ecosystem continues adopting MCP-based AI assistants, adoption may grow modestly (200–500 additional stars annually). Risk of stagnation if primary maintainer deprioritizes or if competing commercial tools integrate similar MCP servers. Unlikely to reach mainstream adoption outside Korean legal domain.
Newsletter
Get analyses like this every Monday
Free weekly digest of the most interesting open-source discoveries.
Languages
No language breakdown available.
Information
- Language
- TypeScript
- License
- MIT
- Last updated
- 2d ago
- Created
- 7mo ago
- Analyzed with
- anthropic/claude-haiku-4-5
Stars over time
No commit data available.
Contributors over time
Top 100 contributors only — repos with more will plateau at 100.
Open issues
No open issues — clean slate.
Open pull requests
No open pull requests.
Top contributors
Contributor data not available yet.
Recent releases
No releases published yet.
Similar repos
saidsurucu/yargi-mcp
Yargı MCP is a FastMCP server providing programmatic access to Turkish legal...
Mcp-Brasil/mcp-brasil
MCP Server that exposes 70 Brazilian public APIs (533 tools across 15 thematic...
yzfly/Awesome-MCP-ZH
A comprehensive Chinese-language curated collection of MCP (Model Context...
| Repository | Stars | Week Δ | Language | Score | Updated |
|---|---|---|---|---|---|
|
chrisryugj/korean-law-mcp
this
|
2.2k | +65 | TypeScript | 8/10 | 2d ago |
|
saidsurucu/yargi-mcp
|
1k | — | Python | 7/10 | 5d ago |
|
Mcp-Brasil/mcp-brasil
|
1.7k | — | Python | 8/10 | 2mo ago |
|
yzfly/Awesome-MCP-ZH
|
7.4k | — | — | 7/10 | 2w ago |
 MicrosoftDocs/mcp
|
1.8k | — | TypeScript | 8/10 | 3w ago |
 ref-tools/ref-tools-mcp
|
1.1k | — | TypeScript | 8/10 | 2w ago |
Reference implementation/documentation for MCP protocol itself; korean-law-mcp is a domain-specific MCP server consuming the protocol, not a competing implementation.
Appears to be an MCP client/framework; korean-law-mcp is a specialized server. Different layer of the stack.
Likely similar regional legal API integration (Brazil), serving analogous niche in Portuguese-speaking market; both are domain-localized MCP servers with comparable adoption patterns.
korean-law-mcp provides AI-native, API-first access to government sources at no subscription cost; lacks the curated editorial content and legal reasoning tools of commercial platforms, but fills the gap for LLM-integrated workflows.
likely precursor point-solutions existed; korean-law-mcp consolidates and extends them under MCP standard, enabling broader client compatibility (Claude, Cursor, Windsurf, etc.) and reducing tool-listing bloat via task parameterization.
