PDFWhisperer pulls the text, geometry and dimensions off a technical sheet and hands back structured data — so the next step, whether that is a model or a person, is working from the page rather than guessing at it.
How extraction works Where the output goesA PDF page becomes a bundle of text items, geometry and parsed dimensions — each with its position on the sheet, not a flat wall of characters.
Drawings that carry no extractable text layer fall through to OCR, so a scan is handled by the same path as a native PDF instead of failing quietly.
The extracted bundle can be put to a language model together with your own notes, so a question about a sheet is answered against that sheet's contents.
Every text item keeps its coordinates and font size, so titles, callouts and schedule cells stay distinguishable from body notes rather than collapsing into one stream.
Vector content on the page is read alongside the text, which is what lets a callout be tied to the thing it points at.
Dimension strings are parsed into values and units instead of being left as loose text for something downstream to re-guess.
Some PDFs use fonts that carry no
ToUnicode map and cannot be decoded at all. Those pages are rasterized
and read by OCR.
The primary output is plain structured data. Anything that can read JSON can consume it.
The same bundle can be rendered into a prompt so a language model reasons over the sheet's real contents.
Extracted geometry can be exported as a starting model for RustyCAD rather than redrawn by hand.
Planned. A bridge for reading DWG files through the same extraction path is designed but not yet shipped.