Taiwan Legal RAG

•
353
•

Provides free semantic retrieval for over 22.5 million Taiwan court judgments, administrative rulings, and Constitutional Court decisions to AI assistants and users.

This tool offers a free, no-signup, no-API-key Taiwan legal Model Context Protocol (MCP) server, allowing AI assistants like Claude, ChatGPT, and Codex to directly perform semantic searches across a vast dataset of over 22.5 million Taiwan court judgments, administrative rulings, and Constitutional Court decisions. It includes robust citation verification mechanisms and provides an open-source command-line interface (CLI) for direct retrieval and bundling of legal documents. Designed to enhance legal AI applications by providing factual retrieval capabilities, it does not generate legal advice or call LLMs, focusing solely on verifiable document retrieval.

Key Features

01Free, No-Signup, No-API-Key Access: Utilize the hosted MCP server or CLI without registration or API keys for public endpoint usage.
02Retrieval-Only MCP Server & CLI: Connects to your existing AI tools (Claude, ChatGPT, etc.) via MCP or CLI, providing contextually relevant legal documents without generating legal advice or calling LLMs.
03Semantic Search across 22.5M+ Taiwan legal documents: Access a comprehensive, structured, and vectorized database of court judgments, administrative rulings, and constitutional court decisions.
04Citation Protection & Verification: Includes `allowed_citations` whitelist, `unread_candidates` flags, and a CLI-based citation checker to combat AI hallucinations of legal citations.
05Case History & Validity Tracking: Automatically appends `case_history` with "廢棄/駁回" flags for judgments and tracks the validity status of administrative rulings.
06298 GitHub stars

Use Cases

01Conducting natural language-based semantic searches for relevant Taiwan court judgments and administrative rulings.
02Enhancing AI assistant capabilities for legal research on Taiwan law by providing verified document context.
03Verifying the factual accuracy of legal citations within AI-generated responses using the bundled document content.