Local MCP assistant for Chinese labor rights and case preparation
worker-rights-cn, from 90le, is an MCP server and AI assistant that brings localized labor-law context to model-driven guidance for mainland China. It supplies AI agents with structured legal texts and tooling to help users calculate statutory entitlements, organize case facts, and produce document drafts for dispute preparation. The design targets mainland Chinese workers, legal aid volunteers, and developers building localized assistance, and it emphasizes local execution to keep sensitive case details inside a user's environment.
What tasks can you actually use the tool for?
The tool connects AI agents to specific legal resources and practical outputs, which supports several concrete tasks. It produces compensation estimates using standard formulas such as N, N+1, and 2N; it formats timelines and evidence into categorized case summaries; and it generates draft arbitration applications, communication letters, and negotiation scripts. Workflows that require structured drafts and numeric entitlement calculations benefit most from these direct, task-oriented outputs.
How accurate are the outputs compared to doing it manually?
Accuracy depends on the model the user pairs with the server and the completeness of case facts supplied. The tool gives models direct access to Chinese labor statutes and arbitration rules, which makes generated answers more context-aware than generic prompts alone. The developer notes it is an information assistant, not a replacement for a licensed lawyer, so users should verify calculations and draft language before filing or signing documents.
Does it require technical knowledge to get useful results?
Using the tool requires an MCP-compatible host and a local runtime: it is built in TypeScript and requires Node.js for execution, and it integrates with MCP clients such as Claude Desktop. Developers and volunteers familiar with local server setup can deploy it directly; non-technical users will typically rely on an intermediary interface or a helper to install and connect the server to an AI agent.
How does the tool handle privacy and workflow fit?
The implementation operates within the MCP framework and is described as privacy-first, it does not automatically save, upload, or transmit personal case materials to external servers. That local execution model suits workflows where confidential facts must remain on a user's machine or organization-controlled host. The project is open-source and aimed at implementers who want auditability and local control over sensitive legal inputs.
Practical preparatory tool for privacy-conscious, technically equipped users
The tool is a practical option for mainland Chinese workers, legal volunteers, and developers who need preparatory drafting and entitlement calculations while keeping data local. Expect to use it as a case-organization and drafting aid rather than a source of definitive legal rulings. Confirm its outputs with a licensed practitioner before formal arbitration or agreement signing, and plan for a brief technical setup to deploy the server.
