PDF Extraction Improvements

We’re improving how Paperguide extracts text, tables, and figures from PDFs—especially messy layouts (multi-column pages, scanned docs, dense tables, and image-heavy papers). These upgrades feed directly into everything downstream, so you’ll see:

• Better answers in Chat with PDF (cleaner grounding + fewer missed sections)

• More reliable inputs for Literature Reviews (less noise, stronger evidence capture)

• Higher-quality extracted tables/data (cleaner rows/columns, fewer broken headers)

• More accurate outputs across all agent workflows that depend on PDF understanding (screening, extraction, synthesis, citations)

In short: cleaner extraction → stronger evidence → better, more consistent AI results.