ALR-TW
Establishes a secure, agentic research harness for Taiwan legal information, integrating official sources with verification and trust gates.
Establishes a secure, agentic research harness for Taiwan legal information, integrating official sources with verification and trust gates.
ALR-TW is a public-preview engineering framework designed to provide a secure and auditable research harness for Taiwan law. It allows external agents and LLMs to initiate research runs via an MCP interface, while maintaining server-side control over data sources, research steps, evidence promotion, and answer verification. Built with a Taiwan civil law perspective, it prioritizes law timeliness, handles court judgments by level, and separates constitutional court opinions. The system integrates with Taiwan Legal RAG (TLR) for enhanced recall of ordinary judgments, but always verifies claims against official sources, making it ideal for developing and deploying legal AI agents with robust safety and accuracy controls.