Source document

Drop a PDF, DOCX, or image file. The extraction and Markdown conversion run on this device.

Drop a document here or choose a PDF or DOCX file (plus images with the OCR engine). Conversion happens locally.
PDF parsing engine
EdgeParse WebAssembly engine. Rust parser with XY-cut reading order and native Markdown tables. Loads a one-time ~2.7 MB module locally on first use.

No file selected. Documents stay on this device.

PDF heading detection