Legalithm
@legalithm
EU AI Act compliance in your editor. Classify risk, cite obligations, draft Article 50 disclosures.
Overview
What is Legalithm?
Legalithm is an offline MCP server that brings EU AI Act (Regulation (EU) 2024/1689) compliance into Cursor and Claude Code. It reads dependency names only, requires no API key or network, and is designed for developers needing a cited, risk-based assessment tool.
How to use Legalithm?
Install by running npx -y legalithm-mcp-server, then invoke the provided tools: classify (risk tier), explain_obligation (obligations for a role+tier), generate_disclosure (Article 50 transparency snippet), and check_record (compliance record by slug).
Key features of Legalithm
- Offline risk classification (unacceptable / high / limited / minimal)
- Cites specific EU AI Act articles for each tier
- Explains obligations by role and risk level
- Generates transparency disclosures in English and German
- Flags low-confidence results for human review
Use cases of Legalithm
- Assess AI tool dependencies for EU AI Act risk tier
- Generate Article 50 transparency statements for AI systems
- Look up published compliance records by slug
- Understand obligations for a specific role and risk level
FAQ from Legalithm
Is Legalithm a substitute for legal advice?
No—it is a cited starting point that flags low-confidence results for human review and explicitly states it is not legal advice.
Does Legalithm require an internet connection?
No—it works entirely offline by reading dependency names only, with no network calls or API key required.
Which editors does Legalithm support?
Legalithm is built for use in Cursor and Claude Code environments.
What data does Legalithm read from my project?
It reads only dependency names; no source code, network traffic, or API keys are accessed.
How can I retrieve a published compliance record?
Use the check_record tool with a specific slug to fetch a published compliance record by name.
