About

What is sens.legal?

sens.legal is an ecosystem in development for traceable legal infrastructure in Brazil. Instead of collapsing everything into a single product, it separates responsibilities by source of authority: jurisprudence, legislation, doctrine, and user experience.

The platform is built around Silo, which integrates a core engine for jurisprudence and reasoning, a frontend interface for lawyers, a legislative resolution component for document-first legislation, and a doctrine pipeline for internal artifact supply.

Who

sens.legal is created by Diego Sens, a lawyer (OAB/PR 96.036) and software engineer building the tools he wished he had as a legal professional. The combination of domain expertise and engineering means every design decision comes from real practice — not an outside interpretation of the field.

Vision

Make legal research, grounding, argumentation, and verification accessible to lawyers and LLMs without giving up traceability. Every claim should point to a real source; every answer should remain explainable.

What makes it different

• Silo with clear component roles — the institutional portal describes Silo's components, each with a clear role and integration contract: the core engine, the interface, legislative resolution, and the doctrine pipeline.
• Core engine as central backend — all legal backend logic lives in the core engine, which owns retrieval, reasoning, verification, chat, and all API/MCP surfaces (50+ REST endpoints, 35+ MCP tools, 1,900+ tests).
• Grounding by authority source — legislative resolution is responsible for what statutes say, the doctrine pipeline prepares internal doctrinal artifacts, and the core engine organizes jurisprudence and reasoning over those inputs.
• Canonical integration surfaces — REST API and MCP live in the canonical docs, avoiding stale manual inventories in the portal.

Current scale

• 2.2M classified court decisions across STJ, TJPR, TJSP, and TRF4
• 75K+ nodes and 120K+ edges in the Neo4j knowledge graph (multi-tribunal production)
• 50+ REST endpoints and 35+ MCP tools
• 1,900+ automated tests across the backend
• OAuth 2.0 with PKCE, 3-stage LLM chat pipeline in production, and ChatGPT App Directory submission
• Accepted into the Neo4j for Startups program ($16K AuraDB Pro credits)

Contact

• GitHub: @sensdiego
• X: @sensdiego
• Email: diegomeyersens@gmail.com