01
Check whether the source contains text
A PDF with selectable words can provide extracted text. A scanned PDF may contain only pictures and return little useful content; use Scan to text for recognition instead. DOCX text is extracted from the document, while HTML scripts and style content are removed from the readable-text route.
02
Inspect boundaries rather than appearance
Use a short sample with a title, two columns, and a table row if those features appear in your real document. Compare the output order to the source. A text extractor cannot always express the spatial relationship between separate columns or cells. Missing spaces or joined paragraphs are editing problems to resolve before reusing the extracted wording.
03
Keep data interpretation explicit
For spreadsheets, inspect how sheet names and cell values are represented, and keep the workbook if formulas or structure matter. For Markdown and JSON, plain text may retain their literal notation. An output file opening successfully says nothing about whether a downstream importer expects those symbols, separators, or line endings.