Legal AIPrivate source
FAUS Legal
Research tooling for Argentine law, with the limits written down.
AI-assisted legal research and case organization for Argentine legal work.
Private-source research product. Public claims are limited to the documented runtime and product scope.
Checked version 2.1.0
Evidence checked 2026-10-30
The problem
Legal research produces long documents, tangled citations and a lot of context that has to stay attached to the right case. A generic chat window loses that structure quickly.
What I built
- Built a research workspace with a supervised web and worker runtime.
- Added document ingestion and retrieval flows for Argentine legal material.
- Kept the product as an aid: it organizes research and does not present itself as a lawyer or legal advice.
Architecture and trade-offs
Web and worker are separate processes
The launcher supervises a web process and a worker process, tracks children, and shuts them down gracefully. It does not rely on pkill or killall.
Local data, explicit operations
The application runs against MongoDB with documented ingestion, data model, operations and security notes. Research state is stored deliberately instead of being reconstructed from a chat transcript.