What problem does it solve? Reading PDFs into an LLM or RAG pipeline is unreliable when files are scanned or have broken text layers, and most converters either lose heading/table structure or require paid cloud OCR that uploads your documents. This Skill converts any PDF to clean Markdown on Windows using Firecrawl's pdf-inspector, with fully local OCR for scanned pages. ## Core Features & Use Cases - Automatic PDF classification and conversion: Detects text-based, scanned, image-based, or mixed PDFs and extracts Markdown with heading, table, and reading-order awareness. - Local OCR for scanned pages: Uses PP-OCRv6 with pinned PDFium and ONNX Runtime libraries so scanned pages are recovered offline without sending documents to any cloud service. - Batch and selective processing: Convert entire folders recursively, specific page ranges, or get structured JSON output with OCR provenance and layout diagnostics. - Use Case: You receive a scanned contract PDF on a Windows machine and need its text for summarization. Run the wrapper script and get a UTF-8 Markdown file with the scanned pages OCR'd locally, plus a report of any low-confidence pages. ## Quick Start Ask the AI to convert a PDF file to Markdown, for example: "Convert C:\docs\report.pdf to Markdown using the pdf-inspector skill and tell me if any pages needed OCR."
