PDFMux

Universal PDF extraction orchestrator that routes pages to the best backend with quality auditing and confidence scoring.

PDFMux is an intelligent PDF text extraction orchestrator that routes pages to the most appropriate backend based on content type. It supports PyMuPDF, OpenDataLoader, RapidOCR, Docling, and Surya backends with cost-aware economy, balanced, and premium processing modes. Features schema-guided extraction for invoices and contracts, RAG-ready chunking, and confidence scoring without requiring a GPU.