calibre DjVu Conversion Text Missing: How to Fix It

When a DjVu file looks normal but a calibre conversion produces blank pages, missing text, an image-only result, or a book that cannot be searched, the problem is usually not an EPUB setting or a broken library. calibre can convert a DjVu document only when the source contains an embedded text layer, typically created through optical character recognition, or OCR. The page images you can see are separate from that hidden text. If the layer is absent, incomplete, incorrectly encoded, or limited to certain pages, calibre has little or no text to place in the converted book. The steps below help you identify that condition, distinguish it from viewer and conversion problems, and choose the safest fix.

1. Confirm the Symptom With a Small Safe Test

Start by testing the source rather than repeatedly changing conversion options. This prevents unrelated settings, device profiles, plugins, or library maintenance from distracting you from the most likely cause.

1.1 Test whether the DjVu contains selectable text

Open the original DjVu file in a viewer that properly supports DjVu. Do not rely only on a thumbnail, browser preview, or image preview. Navigate to a page containing ordinary body text and try to select a sentence with the mouse. Copy it and paste it into a plain-text editor.

Interpret the result carefully:

• If meaningful text pastes correctly, the tested page has an accessible text layer.
• If nothing can be selected, the page may be image-only.
• If selection works but pasted text is empty, scrambled, or unrelated, the OCR layer is defective.
• If some pages work and others do not, the document probably has incomplete OCR coverage.

Search for a distinctive word that appears on the tested page. A working search result is another strong indication that the page has embedded text, although copy and paste remains useful for checking the text's actual quality.

Success looks like this: you can select a sentence, paste readable words into a text editor, and find a visible word through the DjVu viewer's search function. Once this works on several representative pages, stop changing unrelated library, metadata, server, email, or device settings.

1.2 Convert a short representative document

If possible, test a small DjVu file with a few pages or create a copy containing a limited page range using an appropriate DjVu tool. Use material you are authorized to process. Add that file to calibre and convert it with ordinary defaults, first to EPUB or TXT if available for your workflow.

A small test makes the job log easier to inspect and separates a source-format problem from resource limits affecting a very large book. Test pages should include normal paragraphs, not just a cover, title page, diagram, or photographic plate.

Success looks like this: the output contains searchable, selectable text corresponding to the source's hidden text. Formatting may be simpler than the page image because conversion extracts text rather than reproducing the scanned page exactly.

2. Check the DjVu Text Layer and Relevant calibre Options

The defining limitation is straightforward: calibre's DjVu input support expects embedded text. DjVu pages can contain scanned images without any OCR data, and those pages may display perfectly even though there is no text available for conversion. calibre is a converter, not an OCR engine for image-only DjVu pages.

2.1 Understand OCR-generated DjVu files

An OCR-generated DjVu normally stores page images alongside a hidden text layer. That layer enables selection, copying, indexing, and search. It can also record the location of recognized words on each page. calibre reads the embedded text during conversion and uses it to build the output document.

The presence of visible printed words does not prove that this layer exists. A scanned page is still an image from the computer's perspective. Likewise, a filename containing terms such as “OCR,” “searchable,” or “text” is not reliable evidence. Test selection and copying inside the actual file.

OCR quality also matters. calibre cannot reconstruct letters that the OCR engine omitted or correct every recognition mistake automatically. If the hidden layer says “modem” where the page image says “modern,” the converted output may contain the same error.

2.2 Check more than one page

Some DjVu documents combine files created in different ways. The cover may be image-only, front matter may lack OCR, and body chapters may contain text. Other documents lose the text layer on isolated pages during assembly or editing.

Test at least these locations:

• A normal page near the beginning
• A body-text page near the middle
• A page near the end
• Any page that disappears or becomes blank after conversion

If text selection fails only on the missing pages, the diagnosis is complete. Repeated conversion cannot recover text that is not embedded there.

2.3 Review the DjVu input encoding only when text exists

calibre exposes an input-character-encoding option for DjVu conversion. This can help when a document contains embedded text but declares no useful encoding or declares the wrong one. It does not create OCR and will not repair image-only pages.

Consider an encoding change only when copied source text exists but the converted result shows incorrect accented characters, symbols, or scripts. Change one setting, reconvert the small test, and compare the same paragraph. Return to the default if the change makes no improvement.

Success looks like this: characters that were previously garbled become readable while the document's words and paragraphs remain present. If the output is entirely blank, stop experimenting with encodings and return to the embedded-text test.

2.4 Reset saved conversion settings

calibre can remember conversion choices for an individual book. If a previous conversion used aggressive search-and-replace rules, structure detection, or other transformations, those saved settings may affect later jobs.

Open the conversion dialog for the test book and restore the conversion settings to their defaults where appropriate. Pay particular attention to search-and-replace rules because a broad expression can remove text from the intermediate document. Reconvert without custom transformations.

Success looks like this: text returns after the custom rule or saved option is removed. At that point, stop changing global calibre preferences. The cause was the book's conversion configuration, not the DjVu decoder, library, or operating system.

2.5 Separate text extraction from output expectations

EPUB is a reflowable format. A successful DjVu-to-EPUB conversion normally uses extracted text and may not preserve the exact scanned-page appearance, columns, line breaks, diagrams, or coordinates. A PDF output can preserve a page-like presentation in some workflows, but choosing PDF does not cause calibre to OCR missing text.

If your goal is a searchable, reflowable EPUB, the source needs usable embedded text. If your goal is a visual facsimile of image-only pages, direct DjVu conversion through calibre may not provide the result you expect. An OCR or document-processing application should be used before calibre, with an output format suited to the next step.

3. Rule Out Viewer, File, Plugin, and Operating System Problems

Settings for metadata downloads, email delivery, the Content server, and connected devices do not normally determine whether calibre can extract DjVu text. Test the converted file locally before investigating those systems.

3.1 Verify the correct source format

A calibre book record can contain several formats. Confirm that the conversion dialog lists DJVU as the input format. If the record also contains PDF, EPUB, or another format, calibre may use a different source than expected if you select it in the conversion dialog.

Metadata such as title, author, tags, identifiers, and cover art does not supply the book's body text. Refreshing metadata cannot restore a missing OCR layer. Library database maintenance is also unlikely to help when the original DjVu itself lacks selectable text.

3.2 Test the output outside the usual viewer

Open the converted EPUB or PDF in calibre's viewer, then test it in another compatible reader if the text seems missing. Try selecting text, searching for a known phrase, and moving through the document using its table of contents or page controls.

If search works but words are not visible, the output may have a display or styling problem rather than absent text. For example, source styling could theoretically produce poor contrast or unusual placement. If neither search nor selection finds anything, extraction probably failed earlier.

Do not use the Edit book tool to diagnose the original DjVu directly. Instead, use it on a resulting EPUB or another editable output. If the editor's file browser shows HTML documents containing paragraphs, text reached the output and the remaining issue concerns styling or rendering. If the HTML documents contain no body text, investigate the source text layer and conversion log.

3.3 Temporarily disable third-party conversion plugins

calibre uses plugins for many built-in and optional functions. A third-party plugin that intercepts imports, modifies files, or participates in conversion can complicate testing. Temporarily disable only relevant third-party plugins, restart calibre, add a fresh copy of the test file, and convert again.

Do not disable built-in components at random. The purpose is to establish whether an optional extension changes the result.

Success looks like this: the clean conversion contains text when the third-party plugin is disabled. Re-enable other plugins one at a time if necessary, and stop when you identify the conflicting extension.

3.4 Move the test away from cloud-synced or restricted folders

Copy the source to a simple local folder that your user account can read and write, then add that copy to a temporary calibre library. Avoid testing from a cloud placeholder that has not been fully downloaded, a read-only location, an unreliable network share, or removable media that disconnects during processing.

On Windows, macOS, or Linux, security software and permissions can interfere with temporary files, but these problems more often generate an explicit error than silently remove only DjVu text. Check them when the job fails, cannot open the source, cannot create output, or reports access errors.

USB mode and reader-device limitations matter only after local conversion succeeds. If the file has searchable text on the computer but not on a device, the conversion itself worked. Test a device-supported format, transfer it again, and check whether that reader supports searching inside the chosen format. Firewall and Content server settings are similarly unrelated unless the local file works and only network access fails.

4. Use calibre Logs and Conversion Debug Output

Logs are most useful after the basic text-selection test. They can show whether calibre extracted little or no content, selected an unexpected input format, applied a destructive rule, or encountered a decoding error.

4.1 Read the completed job details

Click the Jobs indicator after conversion finishes. Open the completed conversion job and view its detailed log. Search for references to the input format, input plugin, warnings, tracebacks, encoding, empty content, and the output path.

A completed status does not necessarily mean the source provided useful text. It can mean that calibre successfully processed the limited material it was given. Save the log before repeating the test so you can compare results.

4.2 Save conversion debug output

The conversion dialog includes debugging options that can save output from stages of calibre's conversion pipeline to a folder. Choose a new empty local folder, run the small test, and inspect the generated intermediate files.

If the intermediate XHTML or extracted content is already empty, changing EPUB styling, device profiles, or viewer preferences will not restore it. If readable text exists in the intermediate files but disappears later, inspect transformations such as search-and-replace, heuristic processing, or output styling.

Success looks like this: you identify the stage at which the text disappears. Stop changing settings outside that stage.

4.3 Use calibre-debug only when the normal log is insufficient

Advanced users can start calibre's graphical interface in debug mode with calibre-debug -g. On macOS, calibre command-line programs are located inside the application bundle unless their directory has been added to the shell path. Debug mode can expose plugin errors or unexpected exceptions not obvious in the standard job summary.

This step is optional and does not add OCR. If the original DjVu has no selectable or extractable text, a longer debug log will not change the result.

5. Run a Clean Temporary Test Before Reinstalling

Do not delete your library or reinstall calibre as an early response. Neither action adds a missing text layer to a DjVu source, and broad changes can make the real cause harder to isolate.

5.1 Build a controlled test

1. Copy the DjVu to a fully local folder.
2. Confirm that you are authorized to process it.
3. Test selection, copying, and search on several pages.
4. Create a temporary calibre library in another local folder.
5. Add a fresh copy of the DjVu.
6. Keep conversion settings at defaults.
7. Convert to EPUB and inspect text selection and search.
8. Review the job log and debug output if the source text test passed but conversion did not.

If the clean test succeeds, compare the original book's saved conversion settings and relevant third-party plugins. If it fails and the source has no text layer, stop troubleshooting calibre and process the source with OCR software first.

5.2 Use OCR before conversion when necessary

Choose a reputable OCR tool that supports your language and can produce a searchable document or export recognized text. Review the OCR result before importing it into calibre. Check headings, paragraph order, page breaks, special characters, footnotes, tables, and multi-column pages.

You may then convert a suitable searchable output, such as an OCR-generated PDF, DOCX, HTML, or EPUB, depending on what the OCR application can create reliably. Be aware that scanned PDFs and complex layouts can still convert poorly because reflow requires calibre to infer reading order from page-oriented content.

Success looks like this: text can be selected and searched in the OCR application's output before it enters calibre. Once that condition is met, calibre has actual text to process.

6. Quick Fix Checklist

• Open the original DjVu in a full DjVu viewer.
• Select and copy text from several body pages.
• Search for a visible word in the original file.
• Confirm that DJVU is selected as calibre's input format.
• Reset book-specific conversion settings for the test.
• Remove aggressive search-and-replace rules.
• Convert a small file with default settings.
• Inspect the completed job log.
• Save and inspect conversion debug output when needed.
• Test the result locally before sending it to a device or server.
• Temporarily disable relevant third-party plugins.
• Use a local, writable folder outside cloud synchronization.
• Run OCR before conversion if the DjVu is image-only.
• Stop changing calibre settings when the source lacks embedded text.

7. Frequently Asked Questions

7.1 Why can I see words in the DjVu when calibre finds no text?

You may be looking at scanned page images. Humans see words in the image, but conversion software needs encoded characters in an embedded text layer. Without OCR data, the page is effectively a photograph.

7.2 Can calibre perform OCR on an image-only DjVu?

calibre's DjVu conversion support depends on embedded text and should not be treated as an OCR engine. Run the document through appropriate OCR software first, verify the recognized text, and then import a suitable output into calibre.

7.3 Why are only some pages missing after conversion?

The DjVu may have incomplete OCR. Test selection on the exact missing pages. If those pages are image-only while other pages contain hidden text, calibre can convert the available text but cannot infer the absent portions.

7.4 Will converting to PDF preserve the missing words?

Changing the output format does not manufacture a text layer. A PDF may preserve page images in a different workflow, but searchable or selectable words still require OCR text. If you need a visual copy, use a tool designed for page-image conversion. If you need search and reflow, create accurate OCR first.

7.5 Why does the EPUB contain text but look different from the DjVu?

That is often expected. EPUB normally reflows extracted text to fit different screen sizes. The original DjVu may encode the scanned appearance separately from its hidden OCR text, so columns, exact spacing, headers, and image placement may not transfer cleanly.

7.6 When should I stop troubleshooting calibre?

Stop changing calibre settings when text cannot be selected, copied, searched, or extracted from the original DjVu on the affected pages. At that point, the next step is OCR or obtaining a better source file. Also stop once a default conversion in a temporary library works, because that proves the core converter and source text are functional. You can then focus narrowly on the original book's saved settings or a relevant plugin.