PDF to Word: Simple Document Reconstruction
Rebuild simple single-column PDF text as editable DOCX paragraphs, headings and lists. Review inferred spacing and font substitutes.
Choose a simple text PDF, reconstruct editable structure and compare the DOCX with the source. See supported document types below.
Files are processed in this browser, not uploaded to a service.
Limits and output behavior
Simple reconstruction supports single-column black text, basic headings/emphasis, flat numbered/bulleted lists and mixed page sizes. Fonts use Arial, Times New Roman or Courier New substitutes; margins and paragraph spacing are inferred. Wrapping and pagination can differ. Rotated/right-to-left text, colored text, drawings, tables, images, links and running headers/footers are refused. Neither mode provides OCR or general layout-faithful conversion. Text extraction remains an explicit option with more information loss. Maximum 500,000 extracted characters and 40 MB output.
Instructions and examples
From input to output
- Choose a simple single-column PDF with black selectable text and wait for inspection.
- Keep Simple document reconstruction for editable structure. Enter Pages for a subset, such as 1,3-5; leave blank for all.
- Select Create editable Word file. Inspect the Word content, download the DOCX and compare it with the source in Word.
Settings and result behavior
- Reconstruction refuses unsupported content rather than automatically returning a text-only file. Tables, columns, images, links and running headers/footers are not supported.
- Fonts use Arial, Times New Roman or Courier New substitutes. Margins and spacing are inferred; wrapping and pagination may change.
- Text extraction only is a separate option that loses original layout and list/table structure. Use it only when a reviewed selectable-text copy is useful.
A worked example
- Input
Single-column public meeting notes with a heading, wrapped paragraphs and a numbered list.
- Settings
- Word output: Simple document reconstruction; Pages blank.
- Expected result
Actual editable paragraphs and list items, basic emphasis and source page sizes, using inferred spacing and font substitutes.
Supported input and limits
One PDF; 500,000 extracted characters and 40 MB DOCX maximum. No OCR. Reconstruction supports black, unrotated, left-to-right text and flat lists; unsupported geometry/content is refused.
One PDF up to 20 MB and 100 pages; at most 500,000 extracted characters and 40 MB output. Reconstruction supports black selectable text, basic emphasis, flat lists and page sizes. Tables, columns, images, links, running headers/footers, rotated/right-to-left text and scans are refused. Fonts and margins are inferred. Optional text extraction loses more structure. No OCR or general layout-faithful conversion.
If the result is unexpected
- For scans, obtain a text-based source or use a separate OCR workflow. Mixed scanned content is refused by reconstruction.
- After a refusal, use a converter supporting the document features, or explicitly choose text extraction knowing that images and their text can be omitted.
Check before using the result
Compare every selected page. Check words, identifiers, heading roles, list numbering, emphasis and page boundaries; make an edit and save the Word file before using it as a source document.
Report a problem with this tool. Include a nonsensitive example and your browser details.
Questions about this tool
Which PDF documents can I reconstruct?
Simple single-column PDFs with black selectable text. Reconstruction creates editable paragraphs, headings and flat lists, retains basic emphasis and page sizes, and infers fonts, spacing and margins. Tables, columns, images, links, running headers/footers and scans are unsupported. Text extraction is a separate option with more information loss. Neither mode provides OCR or exact arbitrary-layout conversion.
Practical guides
- Convert Selectable PDF Text to Word: What Is Preserved?
Reconstruct simple PDF text as editable paragraphs, headings and lists. Compare the result and understand unsupported tables, images and OCR.