Built-in Text extractor plugins

Caution

Only one text extractor should be active for a given MIME type at a time. Several of the extractors below intentionally overlap in the MIME types they support (e.g. six different PDF extractors, two Excel extractors, several image OCR extractors) so you can pick the technique that best fits your case - but if more than one is enabled for the same MIME type, which one the application actually uses to extract a given document is unpredictable. Disable every alternative you are not using.

The Administration > Utilities > Plugins screen itself detects this: when more than one enabled Text Extractor plugin claims the same MIME type, it lists that MIME type as a conflict - check it after enabling or disabling any of these.

Name	MIME types	Description
AudioTextExtractor	audio/mpeg audio/x-ogg	Reads the file’s own embedded ID3/audio tags (title, artist, album, and so on) using the entagged library; does not analyze the actual audio content.
AutoCadLinkListTextExtractor	image/vnd.dxf image/vnd.dwg	Extracts the TEXT entities (labels/annotations) embedded in an AutoCAD drawing, using the external CADViewer LinkList tool (cadviewer.linklist configuration parameter).
AzureTextExtractor	image/tiff image/gif image/jpeg image/png	OCR via Azure Cognitive Services’ Read API (cloud-based, requires Azure endpoint/key configuration). A cloud alternative to TesseractTextExtractor for the same image types.
EPUBTextExtractor	application/epub+zip	Extracts the plain text of an EPUB e-book’s chapters using epublib.
EmlTextExtractor	message/rfc822	Extracts the headers of an EML mail (From, To, Cc, Date and Subject, one “Name: value” line each) followed by the body, preferring the plain text part over the HTML one. Attachments are skipped, except attached mails.
ExifTextExtractor	image/jpeg	Extracts EXIF/metadata tags (camera model, date, GPS, and so on) from a JPEG using the metadata-extractor library. Not OCR - it does not read any text visible in the image itself.
HTMLTextExtractor	text/html	Strips HTML markup and returns the page’s visible text, using Jsoup.
ImageOpenAITextExtractor	image/tiff image/gif image/jpeg image/png	Sends the image to the configured OpenAI vision model and returns its response as Markdown. An AI-based alternative to TesseractTextExtractor/AzureTextExtractor for the same image types.
JsonTextExtractor	application/json	Extracts every value of a JSON document as a “path/to/field: value” line, so each value keeps its context. Escaped characters (\uXXXX) are decoded; a malformed file is indexed as raw text.
LibreOfficeTextExtractor	application/vnd.oasis.opendocument.text application/vnd.oasis.opendocument.text-web application/vnd.oasis.opendocument.text-template application/vnd.oasis.opendocument.presentation application/vnd.oasis.opendocument.presentation-template application/vnd.oasis.opendocument.spreadsheet application/vnd.oasis.opendocument.spreadsheet-template application/vnd.oasis.opendocument.graphics application/vnd.oasis.opendocument.graphics-template application/vnd.oasis.opendocument.database application/vnd.oasis.opendocument.chart application/vnd.oasis.opendocument.text-master application/vnd.sun.xml.writer application/vnd.sun.xml.calc application/vnd.sun.xml.impress application/vnd.sun.xml.draw application/vnd.sun.xml.writer.global application/vnd.sun.xml.calc.template application/vnd.sun.xml.impress.template application/vnd.sun.xml.draw.template application/vnd.sun.xml.writer.template	Reads the document’s own content.xml directly via SAX parsing - OpenDocument and legacy OpenOffice.org files are ZIP archives of XML - without invoking LibreOffice itself.
MarkdownTextExtractor	text/markdown	Returns the raw Markdown source as plain text (no rendering).
MsExcelTextExtractor	application/vnd.ms-excel application/msexcel application/excel	Extracts cell text from a legacy (.xls) Excel workbook using Apache POI (HSSF).
MsOffice2007TextExtractor	application/vnd.openxmlformats-officedocument.wordprocessingml.document application/vnd.openxmlformats-officedocument.wordprocessingml.template application/vnd.openxmlformats-officedocument.presentationml.presentation application/vnd.openxmlformats-officedocument.presentationml.template application/vnd.openxmlformats-officedocument.presentationml.slideshow application/vnd.openxmlformats-officedocument.spreadsheetml.sheet application/vnd.openxmlformats-officedocument.spreadsheetml.template application/vnd.ms-excel.sheet.macroEnabled.12 application/vnd.ms-excel.template.macroEnabled.12	Extracts text from modern (2007+) Word/PowerPoint/Excel documents, including macro-enabled workbooks, using Apache POI’s OOXML support.
MsOfficeOcrTextExtractor	application/msword application/vnd.openxmlformats-officedocument.wordprocessingml.document application/vnd.openxmlformats-officedocument.wordprocessingml.template application/vnd.ms-powerpoint application/vnd.openxmlformats-officedocument.presentationml.presentation application/vnd.openxmlformats-officedocument.presentationml.template application/vnd.openxmlformats-officedocument.presentationml.slideshow application/vnd.ms-excel application/vnd.openxmlformats-officedocument.spreadsheetml.sheet application/vnd.openxmlformats-officedocument.spreadsheetml.template application/vnd.ms-excel.sheet.macroEnabled.12 application/vnd.ms-excel.template.macroEnabled.12	Converts the document to PDF and OCRs every rendered page with Tesseract, ignoring any native text - useful for scanned or image-heavy Office documents, at the cost of speed. Covers the same Word/PowerPoint/Excel MIME types (legacy and 2007+) as MsWordTextExtractor, MsOffice2007TextExtractor, MsExcelTextExtractor, and MsPowerPointTextExtractor combined.
MsOutlookTextExtractor	application/vnd.ms-outlook	Extracts the subject, body, and attachment names of an Outlook message (.msg) using Apache POI (HSMF).
MsPowerPointTextExtractor	application/vnd.ms-powerpoint application/mspowerpoint application/powerpoint	Extracts slide text from a legacy (.ppt) PowerPoint presentation using Apache POI (HSLF).
MsVisioTextExtractor	application/vnd.ms-visio.drawing	Converts the Visio drawing to PDF via LibreOffice, then delegates to whichever PDF text extractor is currently registered and active to extract its text.
MsWordTextExtractor	application/vnd.ms-word application/msword	Extracts text from a legacy (.doc) Word document using Apache POI (HWPF).
NativeMsExcelTextExtractor	application/vnd.ms-excel application/msexcel application/excel	Alternative to MsExcelTextExtractor using the external catdoc xls2csv command-line tool (system.catdoc.xls2csv configuration parameter) instead of Apache POI.
PdfLayerTextExtractor	application/pdf	Extracts only the PDF’s own text layer (PDFBox); no OCR fallback, so a scanned or image-only PDF returns no text. The lightest and fastest of the PDF extractors.
PdfNewTextExtractor	application/pdf	Same approach as PdfTextExtractor - text layer plus an OCR fallback (Tesseract, on the PDF’s embedded images) when there is no text layer or system.pdf.force.ocr is set - rewritten against the newer PDFBox loading API, with a simpler, Java-only image extraction step.
PdfOcrTextExtractor	application/pdf	Always OCRs every rendered page with Tesseract, ignoring any existing text layer.
PdfOpenAITextExtractor	application/pdf	Always sends every rendered page to the configured OpenAI vision model and returns Markdown, ignoring any existing text layer.
PdfRenderTextExtractor	application/pdf	Text layer plus OCR fallback like PdfTextExtractor/PdfNewTextExtractor, but renders each full page as a 300 DPI image for OCR instead of only extracting embedded image objects - catches more scanned/complex-layout cases, at the cost of being slower.
PdfTextExtractor	application/pdf	Text layer (PDFBox) plus an OCR fallback (Tesseract, on the PDF’s embedded images) when there is no text layer or system.pdf.force.ocr is set; can use the external pdfimages tool (system.pdfimages) instead of Java-based image extraction.
PlainTextExtractor	text/plain text/csv	Returns the raw text content, decoded with the given (or UTF-8) encoding.
PngTextExtractor	image/png image/apng image/mng	Reads the file’s own embedded tEXt metadata chunks (PNG/APNG/MNG binary chunk format). Not OCR - it does not read any visible text drawn in the image.
RTFTextExtractor	application/rtf text/rtf	Extracts plain text from an RTF document using Java Swing’s RTFEditorKit.
RarTextExtractor	application/x-rar-compressed application/vnd.rar	Same as ZipTextExtractor for RAR archives, both RAR4 and RAR5 (including solid archives): the path of every file and, below it, the text of the files that have an active text extractor, with the same limits. Encrypted entries are only indexed by name; when the file names themselves are encrypted, the extraction fails with “Encrypted RAR archive”.
SourceCodeTextExtractor	text/x-java text/x-python text/css text/x-sql application/x-php application/javascript application/x-shellscript	Returns the raw source code as plain text.
TesseractTextExtractor	image/tiff image/gif image/jpeg image/png	Local OCR via Tesseract (system.ocr configuration parameter); optionally retries at extra rotation angles (system.ocr.rotate) to improve results on rotated scans.
XMLTextExtractor	text/xml application/xml application/vnd.scribus	Extracts the text content and attribute values of an XML document via SAX parsing; also handles Scribus project files, which are XML-based.
ZipTextExtractor	application/zip	Extracts the path of every file in the ZIP archive and, below it, the text of the files that have an active text extractor (PDF, Word, plain text, and so on). Nested archives, encrypted entries, images, audio and video are only indexed by name. To protect against oversized archives (zip bombs), at most 1000 entries are read, 20 MB uncompressed per entry and 200 MB in total, and the extracted text is capped at 10 MB; beyond those limits, only names are indexed.

Note

See Creating your own Text Extractor for the interface and a full example.