Reference
Only parts the inspector reads are listed. Each link goes to the ECMA-376 5th edition page. Parts we do not understand still appear in the diagram, labelled as unhandled.
| Part | What we read | Spec |
|---|---|---|
[Content_Types].xml | Default and override content types for every part | ECMA-376 Part 2, §10.1.2 |
_rels/.rels and *.rels | Package and part relationships, including the thumbnail relationship | ECMA-376 Part 2, §9.2 / §10.2 |
docProps/core.xml | dc:creator, cp:lastModifiedBy, title, revision, created, modified | ECMA-376 Part 2, §11.1 (Dublin Core) |
docProps/app.xml | Application, company, template, pages / words | ECMA-376 Part 1, §22.2 extended properties |
docProps/thumbnail.jpeg | Embedded preview, plus JPEG APP1 EXIF text | ECMA-376 Part 2, thumbnail relationship |
word/document.xml | w:ins / w:del tracked changes, comment markers, visible text | ECMA-376 Part 1, §17.13 revisions |
word/comments.xml | Comment author, date, and text | ECMA-376 Part 1, §17.13.4 |
word/people.xml | Presence authors and user ids | ECMA-376 Part 1, people part |
word/styles.xml | Style ids and names (annotation only) | ECMA-376 Part 1, §17.7 |
word/media/* | Embedded images; JPEG EXIF and PNG tEXt / iTXt | ECMA-376 Part 1, §15.2.14 images |
xl/workbook.xml | Sheet names and state="hidden" / veryHidden | ECMA-376 Part 1, §18.2 workbook |
ppt/presentation.xml | Slides with show="0" | ECMA-376 Part 1, §19.2 presentation |
How it works
1
Drop the package
A .docx, .xlsx or .pptx is a ZIP. The page reads the central directory and inflates each part with the browser's DecompressionStream. Stored and deflate only — the methods Office writes.
2
Read the leak report
Core properties, tracked insertions and deletions, comments, people.xml, thumbnail parts, JPEG EXIF, PNG text, and hidden sheets or slides are flagged from the XML itself.
3
Open any part
Click a name in the diagram. You get the raw XML, or a hex preview if the part is binary, plus a plain-English note on what it does and whether it is safe to send.
Why this exists
People are generating Office files from plain text again. Docscope rides the trending repos wordcraft, gridcraft and deckcraft: builders starring those projects need to see what the package they just wrote actually contains before it is emailed.
Plain text in, .docx out. Docscope is the check on the other side: creator, last-modified-by, comments and tracked-change remnants in the file that left the tool.
Spreadsheets from structured text. Hidden sheets (state="hidden") and workbook metadata are exactly what a generated .xlsx can still carry.
Decks from text. A generated .pptx still has a thumbnail, author names, and slides with show="0" that a client can recover.
Pricing
Free
$0
Anyone checking a file before it leaves their machine.
- Full local dissection of .docx, .xlsx and .pptx
- Interactive exploded diagram of every OOXML part
- Leak report: authors, tracked insertions and deletions, comments, thumbnails, EXIF, hidden sheets and slides
- Raw XML with a plain-English note and a safe-to-send verdict
- Works offline once the page has loaded
Pro (founding)
$5/mo founding price — nothing charged today
The consultant or lawyer who sends client files every week.
- One-click scrub that rewrites the file clean
- Batch scanning of whole folders
- Exportable leak report (PDF/HTML) for engagement records
- Founding price locked for early-access members
Team/CI (founding)
$19/mo per repo — nothing charged today
Dev teams and firms that want leaks blocked, not just found.
- Pre-commit hook and CI check that fails on leaky attachments
- Shared scrub policies
- Everything in Pro for the team
Nothing is charged today.
Honest limits
- Does not yet support PDF or DWG files. Office Open XML packages only: .docx, .xlsx, .pptx.
- ZIP method 0 (stored) and method 8 (deflate) only — the methods Office writes. Encrypted packages are not decrypted.
- EXIF reading covers JPEG APP1 text (ASCII tags such as Artist and DateTime) and PNG tEXt/iTXt. It does not decode every TIFF IFD type.
- No one-click scrubbing in the free version. Inspection shows the leak; it does not rewrite the file.
- Batch scanning is only in the Pro version. This page takes one file at a time.
- CI and pre-commit checks are only in the Team/CI version.
- The parsed file list is the ZIP central directory. If a part is missing from that listing, it is missing here too — we do not fill in parts the package does not have.