Docen
← USE CASES·RESEARCH

Scientific PDFs with complex tables & figures

Handle multi-column papers, dense tables, equations, and figure captions without losing structure.

Docen ParseDocen OCRDocen Layout
THE CHALLENGE

Scientific PDFs pack multi-column text, equations, tables, and figures onto a page. Reading order alone is hard; keeping the tables and math faithful is harder.

THE DOCEN APPROACH

Docen recovers reading order across columns, keeps tables as structured cells, preserves equations, and pairs figures with their captions.

  • Restore reading order across columns.
  • Preserve equations and mathematical notation.
  • Keep tables intact as structured cells.
  • Pair figures with their captions.
parse → paper.json
{
  "sections": ["Abstract", "Methods", "Results"],
  "tables": 3,
  "equations": 8,
  "figures": [{ "id": 1, "caption": "Recovery across runs (n = 6)" }],
  "reading_order": "restored"
}
multi-col
reading order
faithful
equations & tables
paired
figures & captions
[]GET STARTED

Try it on yourown documents.

Bring a sample of the documents behind this use case and we'll show you the structured output.