Extraction modes
- Auto Assay
- Lists source choices. Choose a source to run that analysis. Tables remain available even when they contain a single data row.
- Tables
- Uses visible cells and header rows, expanding spans across their covered positions. Nested tables are extracted separately. Values are text.
- Cards
- Ranks repeated groups by size and amount of text, skips typical navigation areas, and lets you choose a group before extraction.
- Links and images
- Exports URLs, visible labels and image metadata. Image exports contain metadata only.
- Pick element
- Click on a sample item to extract it or its repeated siblings. Press Escape to leave the picker.
Records and files
Choose the columns to include. The preview shows up to eight rows. The export uses all rows matching your current selection and filters. Duplicate removal is optional and preserves the original rows while you work.
CSV quotes every field and adds a leading apostrophe to formula-like values, including prefixes concealed by whitespace or invisible characters. Plain negative numbers retain their numeric text. JSON preserves the selected values.
Scan limits
The userscript clips long fields to 2,500 characters and caps scan inputs on large pages. Work in smaller batches when you approach the limits. Browser-only pages and embedded frames are outside its scope.