The rule sets are YAML files a lawyer can read and edit without writing code, and every stage — extraction, risk flagging, version diff, checklist — runs locally against the file you point at.
A Node MCP server for legal document workflows. It extracts and labels clauses from PDF, DOCX and plain text, checks them against YAML rule templates, compares two versions at clause level, and validates a document against compliance checklists. Parsing and analysis run on your machine, with results stored in a local SQLite database.
- Clause boundaries identified and labelled — termination, liability, IP, governing law and more
- Risky provisions flagged against YAML rule templates, returned ranked by severity
- Two versions compared clause by clause, rather than by raw text lines
- A document validated against a checklist such as NDA, employment agreement or SaaS terms, pass or fail per requirement
- Plain-language summaries of the terms
- A batch of documents analysed in one run
- The available templates listed, and analysis reports and audit logs exported
Node.js v20.19 or newer and npm; it runs through npx against mcp-legal-doc-analyzer@latest. Custom compliance templates load from a directory you point it at on the launch command. Extraction is pattern matching and heuristics, and low-confidence results are flagged as such — the output is an analysis aid for a qualified professional to review, not legal advice. Database exports carry the document contents, so treat them as confidential.
One command — npx -y mcp-legal-doc-analyzer@latest