Created attachment 34521 [details] triggering file, based on Tika's testWORD_numbered_list.docx While working on SAX parser for XWPF on Tika, I found that we're not extracting comments anchored in runs in Tika or POI with our DOM parser. unit test: public void testCommentsAnchoredInRuns() throws IOException { XWPFDocument doc = XWPFTestDataSamples.openSampleDocument("testCommentAnchoredInRun.docx"); XWPFWordExtractor extractor = new XWPFWordExtractor(doc); assertTrue(extractor.getText().contains("add a list")); }