OpenXML (Office) format (.docx, .docm, .dotx, .dotm, .xlsx, .xlsm, .xltx, .xltm, .pptx, .pptm, .ppsx, .potx)

The OpenXML format handles Microsoft Office Open XML documents: Word (.docx), Excel (.xlsx), and PowerPoint (.pptx). All three share a ZIP-and-XML package, so one filter serves them and auto-detects the variant from the package contents, per ECMA-376 / ISO/IEC 29500.

Extraction is selective: by default the reader pulls the main body text plus the parts that are normally user-visible (document properties, headers and footers, footnotes, comments, and hyperlink text in Word; shared strings in Excel; and speaker notes in PowerPoint) while leaving hidden, structural, or rarely translated content out unless you opt in. A set of style and colour filters lets you include or exclude runs by paragraph/character style or highlight colour. This mirrors Okapi's OpenXML filter, including its default of accepting tracked changes before extraction.

IDopenxml

SourceBuilt-in

Extensions.docx, .docm, .dotx, .dotm, .xlsx, .xlsm, .xltx, .xltm, .pptx, .pptm, .ppsx, .potx

MIME Typesapplication/vnd.openxmlformats-officedocument.wordprocessingml.document, application/vnd.openxmlformats-officedocument.presentationml.presentation, application/vnd.openxmlformats-officedocument.spreadsheetml.sheet

CapabilitiesRead + Write

Parameters

Parameter	Type	Default	Description
`aggressiveCleanup`	boolean	true	Strip revision IDs, proofing errors, and other noise before merging runs.
`automaticallyAcceptRevisions`	boolean	true	Automatically accept tracked changes before extraction. When true (default, matching Okapi), inserted runs are kept and deleted runs are dropped; rows marked with <w:trPr><w:del/> (ECMA-376 §17.13.5.13) are removed entirely; rows marked with <w:trPr><w:ins/> (§17.13.5.16) are kept.
`codeFinderRules`	array		Regex patterns that match inline codes within translatable text.
`complexFieldDefinitionsToExtract`	array		Field instruction prefixes to extract (e.g., "HYPERLINK", "REF").
`excludeColors`	array		Font colors to exclude (hex RGB, e.g., "FF0000" for red).
`excludedColumns`	array		Column letters to exclude (e.g., "A", "C", "AA").
`excludedSheets`	array		Sheet names to exclude from extraction.
`excludeHighlightColors`	array		Highlight colors to exclude (e.g., "yellow", "red").
`excludeStyles`	array		Paragraph/character style names to exclude.
`extractNonTranslatableContent`	boolean	true	If true (default), non-translatable contextual content is surfaced: image/shape alt-text (descr/title on docPr/cNvPr) as content blocks (visible to ingestion/LLM consumers, skipped by machine translation), and PowerPoint/Excel comment text as data parts, instead of being hidden in opaque skeleton. Disable to keep it opaque.
`extractRunFontsInfo`	boolean	false	Emit font metadata as annotations on blocks.
`fontMappings`	object		Font name to script group mapping (e.g., "MS Gothic": "ja").
`ignoreSoftHyphenTag`	boolean	false	Ignore soft hyphen tags in the document.
`includedSlides`	array		If non-empty, only extract these slide numbers (1-based).
`includeHighlightColors`	array		If non-empty, only extract text with these highlight colors.
`includeStyles`	array		If non-empty, only extract text with these styles.
`lineSeparatorReplacement`	string		Replacement string for line separator characters.
`replaceLineSeparator`	boolean	false	Replace Unicode line separator (U+2028) in output.
`replaceNoBreakHyphenTag`	boolean	false	Replace no-break hyphen tags with the non-breaking hyphen character.
`tabAsCharacter`	boolean	false	Treat tab elements as tab characters instead of placeholder spans.
`translateCharts`	boolean	false	Extract strings from embedded charts.
`translateComments`	boolean	false	Extract comment text from Word documents.
`translateDiagrams`	boolean	false	Extract text from SmartArt diagrams.
`translateDocProperties`	boolean	true	Extract title, subject, keywords from document properties.
`translateFootnotes`	boolean	true	Extract footnotes and endnotes from Word documents.
`translateHeadersFooters`	boolean	true	Extract text from headers and footers in Word documents.
`translateHiddenSlides`	boolean	false	Extract content from hidden slides in PowerPoint.
`translateHiddenText`	boolean	false	Extract text with the vanish (hidden) property in Word documents.
`translateHyperlinks`	boolean	true	Extract hyperlink text for translation.
`translateSharedStrings`	boolean	true	Extract shared strings from Excel workbooks.
`translateSheetNames`	boolean	false	Extract sheet names in Excel workbooks.
`translateSlideMasters`	boolean	false	Extract text from slide masters in PowerPoint.
`translateSlideNotes`	boolean	true	Extract speaker notes in PowerPoint presentations.
`useCodeFinder`	boolean	false	Enable pattern-based detection of inline codes (placeholders, tags, etc.).

Configure it live

Configuration

Loading form…

Examples

Translate everything in a Word document

Include comments, hidden text, and slide masters that are off by default.

translateHiddenText: true
translateComments: true

Limit an Excel workbook to selected sheets

Skip data and lookup sheets during extraction.

excludedSheets:
  - Data
  - Lookups

Extract only specific PowerPoint slides

Translate slides 1, 2, and 5 and skip the rest.

includedSlides:
  - 1
  - 2
  - 5

Processing notes

The document variant (Word / Excel / PowerPoint) is auto-detected from the package contents.
The writer is faithful by default, preserving source run properties inline rather than rewriting them into synthesised styles.

Limitations

One filter covers Word, Excel, and PowerPoint; some options apply only to the relevant variant (for example, the slide options affect PowerPoint only).

Reference documentation

← Back to the Format Reference