Office embedded image extractor
Document tools
Loading
Loading tool
The tool is loaded only when you open it.
All processing for this tool happens in your browser. Your input is not sent to a server.
About this tool
Inspect transitional DOCX, XLSX or PPTX locally. Follow reachable internal image relationships to direct word/media, xl/media or ppt/media parts. Export PNG/JPEG/GIF/WebP only when the declared content type matches its header. No pixel decoding, image rendering, re-encoding, macro execution or external fetch. This does not prove image validity or full OOXML conformance.
Common uses
- Recover original local raster resources from Word, Excel and PowerPoint packages.
- Review explicit media exclusions without rendering documents or contacting external image URLs.
How to use it
- 1.Select one supported Office package and inspect its image parts.
- 2.Review generated file inventory, relationship counts and all omitted-media/thumbnail counts.
- 3.Acknowledge exclusions if present, then save inventory JSON or the ZIP containing inventory.json and original image bytes.
Synthetic image-part classification
Synthetic image-part classification
{
"[Content_Types].xml": "<Types xmlns=\"http://schemas.openxmlformats.org/package/2006/content-types\"><Default Extension=\"png\" ContentType=\"image/png\"/><Default Extension=\"gif\" ContentType=\"image/gif\"/><Override PartName=\"/word/document.xml\" ContentType=\"application/vnd.openxmlformats-officedocument.wordprocessingml.document.main+xml\"/></Types>",
"_rels/.rels": "<Relationships xmlns=\"http://schemas.openxmlformats.org/package/2006/relationships\"><Relationship Id=\"main\" Type=\"http://schemas.openxmlformats.org/officeDocument/2006/relationships/officeDocument\" Target=\"word/document.xml\"/></Relationships>",
"word/document.xml": "<w:document xmlns:w=\"http://schemas.openxmlformats.org/wordprocessingml/2006/main\"/>",
"word/_rels/document.xml.rels": "<Relationships xmlns=\"http://schemas.openxmlformats.org/package/2006/relationships\"><Relationship Id=\"image\" Type=\"http://schemas.openxmlformats.org/officeDocument/2006/relationships/image\" Target=\"media/demo.png\"/><Relationship Id=\"external\" Type=\"http://schemas.openxmlformats.org/officeDocument/2006/relationships/image\" TargetMode=\"External\" Target=\"https://example.test/never-load.png\"/></Relationships>",
"word/media/demo.png": "89504e470d0a1a0a0000000d494844520000000100000001",
"word/media/unused.gif": "47494638396101000100000000",
"docProps/thumbnail.png": "89504e470d0a1a0a0000000d494844520000000100000001"
}{"candidates":2,"extracted":1,"unsupported":0,"unreferenced":1,"external":1,"previews":1,"nonMedia":0}The fixture has one header-admitted PNG, one unreferenced GIF, one external image relationship and one package thumbnail. No pixels are decoded; these short synthetic headers are a classification fixture, not complete valid pictures.
Limits and notes
- Limits:16 MiB archive,8 MiB per entry,48 MiB expanded,1200 entries;2 MiB per parsed XML,100000 total parsed nodes,depth64;200 media parts,32 MiB copied images;128 KiB inventory. FileReader15s and ZIP unpack15s. Preview12000 UTF-16 units. ZIP≤32 MiB+128 KiB+65536 overhead. Cancel aborts file reading and discards late ZIP results; already-started ZIP workers finish within their limit. XML and ZIP packaging are bounded synchronous operations, not a whole-process memory ceiling.
- Reference counts are image relationship records from parts reachable from the main part, not visible placements or page/slide occurrences. Unreferenced media, unsupported/vector parts, package thumbnails and targets outside the standard media directory are listed or counted separately and are not exported.
- No upload or persistence. Generated filenames and inventory omit source names, document text, XML and relationship targets. Original image bytes may retain private metadata and may represent hidden or unused visual resources. Extraction is not anonymization and does not reproduce page layout.
- Select one nonempty .docx, .xlsx or .pptx no larger than16 MiB. Legacy DOC/XLS/PPT, macro-enabled and encrypted documents are unsupported.
- Inspect transitional DOCX, XLSX or PPTX locally. Follow reachable internal image relationships to direct word/media, xl/media or ppt/media parts. Export PNG/JPEG/GIF/WebP only when the declared content type matches its header. No pixel decoding, image rendering, re-encoding, macro execution or external fetch. This does not prove image validity or full OOXML conformance.
Frequently asked questions
Are these rendered page screenshots?
No. These are original internally related raster parts. Cropping, transforms, compositing, visible occurrence and page/slide layout are not reconstructed. Relationship counts do not prove a picture is visible.
Are SVG or external images downloaded?
No. Only declared PNG/JPEG/GIF/WebP with admitted headers are copied. Vector and other unsupported media are omitted; external targets are never fetched.
Does extraction remove private metadata?
No. Generated names and inventory omit source metadata fields, but original image bytes are unchanged and can contain metadata or private content.