About DJVU Files
DjVu Document
DjVu is a document format designed for scanned documents with high compression.
Family
Documents & Text
Extension
.djvu
MIME Type
image/vnd.djvu
Can Use As
HOW DJVU
CAME TO BE.
DjVu — pronounced "déjà vu" — was developed at AT&T Labs in 1996 by Yann LeCun (later the inventor of modern convolutional neural networks), Léon Bottou, Patrick Haffner, and Paul Howard. The goal was extreme compression of scanned documents: take a 300 DPI color page scan that weighed 25 MB as TIFF, separate it into a foreground (text and line art) and background (photos and paper texture) using wavelets, and compress each layer aggressively. A typical scanned page dropped from 25 MB to under 50 KB while staying perfectly readable.
Internet Archive, Project Gutenberg, and library digitization projects adopted DjVu throughout the 2000s for millions of scanned books. Browser support was always via a plugin (the Caminova/LizardTech one), which meant the format struggled as the plugin era ended. Today DjVu’s niche has mostly been filled by PDF/JBIG2, but you still find it anywhere scanned archives are kept — Wikipedia, Archive.org, and most academic library scans default to DjVu for the ratio.
CURIOSITIES &
TRIVIA.
DjVu was co-invented by Yann LeCun, who later founded Meta AI Research and won the Turing Award for work on convolutional neural networks.
A typical 300 DPI color scan of a book page is 25 MB as TIFF, ~600 KB as PDF, and ~50 KB as DjVu at perceptually equal quality.
Archive.org's Open Library stores millions of scanned books with a DjVu version alongside the PDF.
The trick is layered wavelet compression: text and line art go in a foreground bitplane, photos and paper stain go in a low-res background, both recombine at display time.
Browser plugin support died with NPAPI around 2015 — which is why PDF won the scanned-document war despite being 5-10× larger.
STRENGTHS &
LIMITATIONS.
Strengths
- Extreme compression for scanned documents — typically 5-10× smaller than PDF.
- Preserves OCR text layer alongside scanned image.
- Mature archival format at Internet Archive, Wikipedia, academic libraries.
- Separate foreground/background layers compress each content type optimally.
Limitations
- No native browser support since the NPAPI era (~2015).
- Tooling outside Linux is sparse.
- Encoding requires the proprietary LizardTech/Caminova toolchain or djvulibre.
- PDF/JBIG2 now covers most of DjVu's use cases with wider compatibility.
Typical Sizes & Weights
Scanned page (B&W, text-only)
10-30 KB
Scanned page (color, illustrations)
40-150 KB
300-page scanned book
5-30 MB
Technical Specifications
- MIME type
- image/vnd.djvu
- Extensions
- .djvu, .djv
- Compression
- Wavelet (IW44) + JB2 for line art
- Invented at
- AT&T Labs (1996)
- Reference tool
- djvulibre (open source)
CONVERT FROM
DJVU
Common Use Cases
Scanned document archives
Related Formats
Popular comparisons
DJVU vs PDF
Differences, file size, and when to choose each format.
DJVU vs DOCX
Differences, file size, and when to choose each format.
DJVU vs ODT
Differences, file size, and when to choose each format.
DJVU vs TXT
Differences, file size, and when to choose each format.
DJVU vs MD
Differences, file size, and when to choose each format.
DJVU vs HTML
Differences, file size, and when to choose each format.
Popular DJVU conversions
The most-requested destinations when starting from DJVU.
Frequently Asked Questions about DJVU
Frequently Asked Questions
DJVU (DjVu Document) is a document format used to store paginated text, with optional formatting, tables, images, hyperlinks, headers and footers. It sits in the documents & text family and is typically associated with a specific office suite or publishing pipeline that defined the format and ships the canonical reader.
Modern office suites — Microsoft Word, Google Docs, LibreOffice Writer, Apple Pages — open most DJVU files with reasonable fidelity. If your installed software does not support DJVU, convert to DOCX or PDF first using KaijuConverter; both open in virtually every reader, including free online viewers.
Upload the DJVU to KaijuConverter and pick DOCX, PDF, ODT, RTF, HTML, Markdown, or plain text. Our pipeline runs LibreOffice headlessly plus pandoc for text formats — the same engines behind professional document pipelines. Styles, tables, images, and hyperlinks survive the conversion intact.
Yes, to a high degree. Standard fonts, headings, lists, tables, images, hyperlinks and page structure transfer cleanly. Custom fonts substitute to the closest match if not embedded; obscure layout features unique to DJVU (frames, legacy macros) may flatten to static content in the target format.