What does this document content extractor do?
Open one .docx, .pdf, .pptx, or .epub file in your browser. The tool lists the content it can extract by category, lets you select individual items, and exports the selection as TXT, Markdown, CSV, JSON, or an available ZIP. It processes the file locally and does not modify the original.
Format support
- DOCX includes paragraphs, headings, flattened table rows, numbered or bulleted list items, and embedded media from
word/media/. - PDF includes selectable page text, PDF image references, links, bookmarks/outline, and font metadata: family, page, and vertical orientation. Image items can be previewed even when their raw data cannot be resolved.
- PPTX includes slide text and embedded media from
ppt/media/. Speaker notes are excluded. - EPUB includes chapters, the cover image, and navigation/table of contents. DRM reporting is based on a
META-INF/encryption.xmlfile.
Preview, search and selection
Review the table's Reference, Preview, and Details columns. Category chips filter content types. Search checks each item's reference, short preview, and details; it does not search the complete paragraph, chapter, or page-text payload. Use the checkboxes, Select all in view, and Clear selection to control the export.
Export formats and compatibility
- TXT exports text-oriented items such as paragraphs, headings, slide text, page text, and chapter text.
- Markdown exports headings, lists, paragraphs, and flattened table rows.
- CSV exports DOCX tables or metadata for categories such as media, links, bookmarks, images, and fonts.
- JSON exports structured item metadata, including PDF page text and EPUB navigation.
- ZIP extracts binary media when the selected item has extractable binary data. This applies to supported DOCX/PPTX media and the EPUB cover; selecting a PDF image may produce a failed item because PDF image entries do not have a binary extractor.
Export buttons are calculated from the formats supported by every selected category. A mixed selection can therefore offer fewer formats or show an incompatibility error.
How to extract content from a document
- Drop one DOCX, PDF, PPTX, or EPUB file onto the drop zone, or click browse file.
- Review the stat cards and any warnings - encryption, DRM, or missing text layers are reported here.
- Use the category chips and search to browse extracted items, then select what you want.
- Pick an export format appropriate to the selected content and download the result.
Limitations and local processing
The file is read and processed locally in browser memory with the File API, JSZip, and pdf.js. Generated exports are also created locally, and the original file is not modified.
- There is no OCR. PDF pages without selectable text still appear as items marked likely scanned or image-only.
- DOCX tables are flattened to cell-text rows; merged cells, nested tables, and column widths are not reconstructed.
- Encrypted or password-protected files cannot be decrypted. EPUB DRM reporting covers
META-INF/encryption.xml. - Only the first 2,000 PDF pages are processed. Extraction stops adding items at 20,000 items, so very large files may produce incomplete results.
- Referenced PDF images may not be resolvable for extraction, even when an image reference is listed.
Frequently asked questions
Which file formats can this tool extract from?
DOCX, PDF, PPTX and EPUB. DOCX gives paragraphs, headings, tables, lists and media; PDF gives page text, image references, links, bookmarks and font metadata; PPTX gives slide text and media (not speaker notes); EPUB gives chapters, a cover image and navigation.
Does this tool run OCR on scanned PDFs?
No. It only reads the selectable text layer that already exists in a PDF via pdf.js. If a page has no extractable text layer, it's reported as likely scanned or image-only rather than guessed at or reconstructed.
Does it reconstruct complex table layouts?
No. DOCX tables are flattened to rows of cell text in source order. Merged cells, nested tables and column widths are not reconstructed - this is a content extraction tool, not a layout tool.
Does this tool upload my document anywhere?
No. The file is read and processed entirely in your browser using the File API, JSZip and pdf.js. Nothing is sent to a server, and the original file is never modified or written back to.