What problem does it solve? Judicial processes downloaded from systems like PJe, SEI, or Kanoê arrive as large merged PDFs that are hard to read, search, or feed into analysis workflows. This Skill converts them into clean, structured Markdown, splitting each document (sentença, despacho, petição, etc.) into its own file inside a case-numbered directory. ## Core Features & Use Cases - Automatic document splitting: Detects document boundaries inside merged court PDFs using title markers (SENTENÇA, DECISÃO, DESPACHO, PETIÇÃO INICIAL) and saves each as a separate .md file. - Dual extraction engines: Prefers markitdown for high-fidelity conversion, with a PyMuPDF fallback for offline or credential-free environments; optional Azure Document Intelligence OCR for scanned PDFs. - Case index generation: Produces an INDEX.md with the case number (CNJ/NUP) and a linked table of all extracted documents. - Use Case: You download a 500-page process from PJe and need each peça processual as a separate searchable Markdown file for legal analysis or LLM ingestion. ## Quick Start Ask the agent to convert the attached court process PDF into a folder of Markdown files organized by document type, using the case number as the directory name.
