DOCX Comments Extractor

Output format

Also the file you download

Drop a reviewed .docx hereor click to browse

Runs in your browser
No uploads
Never modifies your file

Reads review comments only — tracked changes, revisions and formatting are left alone. Your document is read in this tab and never written to.

How to export comments from a Word document

Drop the .docx

Drag the reviewed document onto the drop area, or click to choose it. It is read in your browser and never modified.
Read the table

Every comment appears with its author, date, the document text it is anchored to, the comment itself, the comment it replies to, and whether the thread is marked resolved.
Pick a format

CSV opens in Excel, Numbers or Sheets. A Markdown table pastes into a ticket, a pull request or a wiki page.
Download

The file is named after the document, so review-v3.docx gives you review-v3-comments.csv.

Common uses

Turning a review round into a task list

Thirty margin comments become thirty rows with an author and an anchor, which can be sorted, assigned and ticked off instead of re-read.

Answering a legal or compliance markup

A contract that comes back annotated needs every point addressed and recorded. A table of author, clause text and comment is the record.

Summarising feedback from several reviewers

Sorting by author shows who raised what, and which comments are replies to someone else rather than new points.

Reading comments without Word

On a machine with no Word installed, the comments in a .docx are effectively invisible. This reads them out of the file directly.

Why pull comments out of a Word document?

A document comes back from review with thirty comments in the margin, and the next step is never "read them in Word". It is a list: what was raised, by whom, against which sentence, and which ones are already answered. Getting that list today means opening the Review pane and copying comments one at a time into a spreadsheet, which is slow enough that people skip it and work straight from the margin instead. The comments are already structured data inside the file — this page reads them and hands back the table.

DOCX Comments Extractor reads the review comments out of a Word document and presents them as a table you can export. A .docx is a zip, and comments live in their own parts of it: word/comments.xml holds each comment with its author, initials, date and text, word/commentsExtended.xml records which comment is a reply to which and whether a thread has been marked resolved, and the markers in word/document.xml say which passage each comment is anchored to. This page opens the package in your browser and reads all three.

The result carries seven columns: the comment number, the author, the date, the anchored text from the document, the comment body, the comment it replies to, and whether it is resolved. Export as CSV for a spreadsheet, or as a Markdown table for a ticket or a pull request description.

Two honest limits. Tracked changes are not comments — insertions, deletions and formatting revisions live elsewhere in the document and are not extracted here. And the reply and resolved columns depend on the part that Word 2013 and later write; a document from an older version, or from another word processor, still gives up its comments but has no thread information to report.

The document is only ever read. Nothing is written back into the file, and nothing is uploaded — the package is opened in the tab, which you can confirm in your browser's network tab.

How it compares

The manual route is the Review pane: open each comment, copy the text, paste it into a spreadsheet, add the author and the sentence it referred to by hand. It works and it is the reason most review rounds never get a summary at all.

Scripted routes exist — a Python library, a macro — and they are the right answer when this runs nightly over a folder. For the document sitting in front of you they are an install and a runtime.

Upload-and-extract services do the same parse on a server. A reviewed document is close to the worst thing to hand to one: review comments are candid by nature, often about people, and frequently attached to a contract or a manuscript that is not public. Reading the zip locally avoids the question.

It sits next to the other two tools that look inside a .docx: DOCX Metadata Remover, which reports and strips the properties a file carries about its authors, and DOCX to Markdown, which converts the body text and leaves the comments behind.

Tips

The anchored text is the passage the comment points at, read from the range markers Word writes around it. A comment attached to a single word shows that word.
Replies are linked through the paragraph identity Word stores, not through the comment number, so a reply shows the comment it actually answers.
The resolved flag comes from the part Word 2013 and later write. A document produced by an older version, or by another editor, has no resolved information and every row reads as unresolved.
A comment with no author usually came from a file where the reviewer's name was stripped — the DOCX Metadata Remover does exactly that, deliberately.
The CSV is plain RFC 4180 with a header row and no preamble, so it opens in a spreadsheet without an import step.
Tracked changes are not comments. This tool does not extract insertions, deletions or formatting revisions.

Frequently asked questions

Is my document uploaded?

No. The .docx is opened in your browser and the parts inside it are read there. Nothing is sent to a server and nothing is stored.

Does it change my document?

No. The file is read only. No comment is removed, resolved or edited, and no new file is written back — the only download is the comment table.

Does it extract tracked changes too?

No. Comments only. Insertions, deletions and formatting revisions are a separate feature of Word stored elsewhere in the document, and this tool does not read them.

What if the document has no comments?

You get a plain message saying so. That is not an error — a document with nothing in the Review pane is simply a document with no comments part inside it.

How are replies detected?

Word records the thread shape against the identity of each comment's last paragraph rather than against the comment number. The tool follows that link, so a reply is matched to the comment it actually answers.

Why is every comment showing as unresolved?

The resolved flag lives in a part that Word 2013 and later write. If the document came from an older version, or from another editor that does not write it, there is no resolved information in the file to report.

What is the anchored text column?

The passage in the document that the comment points at, taken from the range markers Word puts around it. It is what makes the exported table readable without opening the document alongside it.

Will the CSV open cleanly in Excel?

Yes. It is plain RFC 4180 — a header row then the data, with commas, quotes and line breaks inside a comment quoted properly. There is no banner or comment line above the header to trip up the import.

Can it read an older .doc file?

No. A .doc is a binary format rather than a zip, and it is rejected with a message rather than a broken result. Save it as .docx in Word first.