fileformer
search⌘KtranslateENexpand_moreupload_fileConvert
Converters8 file families · 44 direct conversions
What changes in a conversionarrow_forward
home/document/compressor

Make documents smaller, and see which bytes went.

PDF, Word, Excel, PowerPoint, OpenDocument, EPUB, HTML and 19 more. Each file is rebuilt on your device, part by part, then opened again to check nothing was lost. Nothing is uploaded.

descriptionWord filepicture_as_pdfPhoto PDFcodeWeb page
27 formats, five wayshow each one gets smaller
ZIP repack + images14 formats.docx.docm.dotx.dotm.odt.kwd.sxw.xlsx.ods.pptx.odp.xps.oxps.epub
PDF rewrite1 format.pdf
Whitespace and markup7 formats.rtf.abw.aw.txt.csv.html.md
Convert first4 formats.doc.dot.xls.ppt
Already at its floor1 format.djvu
quality preset · image size limit
Fast72 dpi · 792 pxNormal96 dpi · 1056 pxMaximum150 dpi · 1650 pxUltra300 dpi · 3300 px
Optimize imagesresize photos to the presetRemove metadataauthor, company, edit time
descriptionquarterly-report.docxDOCX · 545 KBSaved without compression, as some exporters write it, with a 2400 × 1600 photo and the author’s name in its properties.Compress this sample →
picture_as_pdfcoastline-brochure.pdfPDF · 1.3 MBTwo full-size photos, uncompressed page text, XMP metadata, and an older photo left behind by an incremental save.Compress this sample →
codeharbour-newsletter.htmlHTML · 16 KBA page exported by a site builder: comments on every block, deep indentation, and a pre block that must survive.Compress this sample →
Or drop up to any number of documents anywhere on this panel.
filebefore → aftersavedresult
0 file · 0 B → 0 B Clear
Where the bytes went in
partbeforeafterwhat changed
By kind of content
Checked after compressing
~/same-report-every-format

How much each format really shrinks

One report written in 13 formats on this page, plus a brochure PDF with two photos, compressed at the preset you picked above. Run it again after changing the preset to see what each step costs.

10/13samples got smaller
−91%best saving
13passed the reopen check
Compressing 0 of 14…hourglass_topMeasuring…
samplemethodbeforeaftersavedcheck
.docxDOCX with a photo, stored uncompressednot measured yet--
.pdfPDF with two photosnot measured yet--
.htmlHTML exportnot measured yet--
Photos are almost all of itIn the PDF and the DOCX nearly every saved byte comes from resizing pictures. With Optimize images off, both shrink far less.
A ZIP can only be squeezed onceThe second row is the same DOCX after one pass, measured with images off: deflating parts that are already deflated gives almost nothing back.
Text is already smallTXT, Markdown, CSV and RTF hold almost no slack. To shrink them further, put them in a ZIP with the Archive Converter.
~/all-27-formats

What happens to each format

The method is chosen from the file’s bytes, not its name, so a renamed file is still handled correctly.

formatwhat it iswhat this page doeswhat to expectmore
Word processing13 formats
. →
.docxWord documentUnpacked, images resized, every part deflated again.Word already deflates its parts, so a text-only file barely moves; photos are where the saving is.What is DOCX →
.docmWord, with macrosRepacked like DOCX. The VBA project part is copied byte for byte.Same as DOCX; macros keep working. →
.dotxWord templateRepacked like DOCX.Styles and layout are untouched. →
.dotmWord template, macrosRepacked like DOCX, macros copied as they are.Same as DOTX. →
.odtOpenDocument textRepacked; mimetype stays first and uncompressed, as ODF requires.LibreOffice already deflates parts; images decide the size. →
.kwdKWordRepacked at the browser’s DEFLATE.Old KOffice files often store parts uncompressed, so they shrink. →
.sxwStarOffice WriterRepacked, mimetype kept first.Opens exactly as before. →
.rtfRich Text FormatLine breaks removed where RTF ignores them. Files with \bin data are left alone.Embedded pictures are hex text and dominate the size; this cannot shrink them. →
.abwAbiWordWhitespace between XML tags removed, then reopened to compare word counts.A few percent. →
.awApplix WordsTrailing spaces and repeated blank lines removed.Small. →
.txtPlain textTrailing spaces and runs of blank lines removed.Usually nothing: text is already minimal. Zip it to go further. →
.docWord 97-2003Not rebuilt here: the binary format has no ZIP layer to repack.Convert to DOCX first; that alone often halves it.DOC to DOCX →
.dotWord 97-2003 templateNot rebuilt here.Convert to DOTX through DOC to DOCX.DOC to DOCX →
Spreadsheets4 formats
. →
.xlsxExcel workbookSheets repacked, images resized.Large repetitive sheets give a few percent. →
.odsOpenDocument spreadsheetRepacked, mimetype kept first.Formulas and formatting untouched. →
.csvComma-separated valuesBlank lines outside quoted cells removed.Almost always nothing to take. →
.xlsExcel 97-2003Not rebuilt here.Convert to XLSX first.XLS to XLSX →
Presentations3 formats
. →
.pptxPowerPointSlides repacked, photos resized to the preset.Photo-heavy decks shrink the most of any Office file. →
.odpOpenDocument presentationRepacked, pictures resized.Same as PPTX. →
.pptPowerPoint 97-2003Not rebuilt here.Convert to PPTX first.PPT to PPTX →
Fixed layout, web and ebook7 formats
. →
.pdfPDFRewritten object by object: JPEG photos resized, uncompressed streams Flate-compressed, leftovers from old saves dropped.The biggest savings of any document format, mostly from photos.What is PDF →
.xpsXPSRepacked, page images resized.Convert to PDF if you need much smaller. →
.oxpsOpenXPSSame as XPS.Same as XPS. →
.epubEPUBRepacked with mimetype first; cover and inline pictures resized.Text-only books gain little; images decide it.What is EPUB →
.htmlHTMLComments and indentation removed; pre, textarea, script and style kept exactly.Exported pages often lose 30-50%. →
.mdMarkdownTrailing spaces trimmed, except the two that mean a line break.Tiny already. →
.djvuDjVuLeft as it is.Scanned pages are already compressed with JB2 and IW44, built for exactly this. →