A .docx Word document stores every external URL as an Office Open XML hyperlink relationship inside its package, which means a complete list of URLs can be read from a single internal XML file rather than scraped from the visible page. To get every URL from a Word document, open the Word Hyperlink Extractor in your browser, select the .docx file from your device, and the tool reads that relationship file locally to return a clean list of http, https, and mailto destinations. The output is a plain-text report with one URL per line that can be downloaded, pasted into a spreadsheet, or handed off to a colleague for review. Because the relationship file is the authoritative source of where each link points, the resulting list does not depend on guessing URLs from visible text or scrolling through formatted paragraphs. The tool deliberately returns only safe external web and email destinations, so internal headings, bookmarks, local file paths, and relative targets are filtered out before the report is shown. Nothing leaves your machine during this process, and the extractor never visits, resolves, redirects, or scores any of the destinations it returns.
What Counts as a URL Inside a Word Document
Word can show many things that look like links, but only a subset of them are external URLs in the strict sense. The Office Open XML format describes a .docx package as a ZIP archive that contains separate XML files for the document body, styles, settings, and relationships. External URLs are recorded in a dedicated file called word/_rels/document.xml.rels, where each relationship carries a Type, an Id, and a Target. Hyperlinks are the relationship type used for external URLs, and the TargetMode is set to External to distinguish them from internal references.
This structure matters because Word can also show other clickable objects that are not external URLs at all:
- Bookmarks and internal cross-references live in the same relationship file but with TargetMode Internal, so they only point to other parts of the same document.
- Citations managed by Word's bibliography tool are stored in a separate XML part and are not hyperlink relationships.
- Buttons, shapes, and embedded objects can carry their own click behavior through a different XML element and are not relationships either.
- Fields such as HYPERLINK, INCLUDEPICTURE, or MERGEFIELD can render visible text that looks like a URL but are stored as field instructions rather than as relationships.
- Plain text that simply looks like a URL, for example a footnote written as "https://example.com" without being a hyperlink, is not represented in the relationship file at all.
A precise URL extractor works against the relationship file directly, which means it returns only what Word itself treats as an external hyperlink. It does not need to render the document, run its field codes, or interpret any of the other clickable objects described above. Word documents often hide their useful destinations behind formatted text, buttons, citations, and images, so trying to copy visible text is slow and easily misses links attached to an image or a short phrase.
Why Pull Just the URLs (and Not the Link Text)
There are several everyday reasons to want only the URL strings from a Word document. Source auditing before publishing a report is one of the most common: a reviewer can paste the TXT list into a URL checker, scan it for outdated domains, and confirm that every citation in the document has a real external destination on record. A quick handoff for another team is another reason. Instead of forwarding a .docx file, you can hand over a single plain-text list of every external URL the document points to, which is easier to triage, comment on, or re-format in another tool. Reference collection from a document received from outside is a third case. When a colleague emails a long Word file full of references, extracting the URLs gives a clean, deduplicated starting point for a bibliography. Editors and accessibility auditors also rely on URL-only lists when reviewing templates for reuse, because they need to know exactly which destinations a template links to before approving it. Finally, pre-flight checks for broken or risky links work well with a plain-text URL list, since it can be run through a separate URL-checking process so reviewers can flag domains without opening Word at all. For each of these tasks, the URL itself is the only piece of data that matters; the visible anchor text, the page number, and the surrounding paragraph all stay in the original document and can be consulted there when context is needed.
Extract URLs from a Word Document Step by Step
For a clean, deduplicated list of the external URLs inside a .docx file, the Word Hyperlink Extractor reads the relationship file directly. The three operating steps are deliberately short so the document never leaves the browser:
- Choose one .docx Word document from your device.
- Wait while the browser reads external hyperlink relationships from the document package.
- Review the safe destinations and download the one-link-per-line TXT report.
To use it in practice, open the page in your current browser tab, click the file picker, and select the .docx you want to inspect. The browser validates the ZIP package, opens word/_rels/document.xml.rels, and keeps only the Relationship entries whose Type is the hyperlink relationship and whose TargetMode is External. XML entities are decoded, http, https, and mailto targets are normalized, and duplicates are removed before the list is rendered on the page. From there, you can either scan the results in the browser or download a plain-text file that contains one URL per line and is ready to paste into a spreadsheet, a URL checker, or any local text workflow.
What the URL Report Includes and Skips
The extractor is deliberately conservative, so it is worth knowing exactly which targets count as a URL for the purposes of this tool. The table below summarizes the categories that are reported versus the categories that are filtered out, based on the documented behavior of the relationship-file reader.
| Target category | Included in the TXT report? | Reason |
|---|---|---|
| http URLs | Yes | Standard external web target |
| https URLs | Yes | Standard external web target |
| mailto URLs | Yes | Treated as a safe external destination |
| Internal headings and bookmarks | No | TargetMode is Internal, not External |
| Relative or local file paths | No | Not an external web destination |
| JavaScript-style or unsafe protocols | No | Excluded as unsafe protocols |
| Malformed or non-URL targets | No | Rejected during XML validation |
| Duplicate destinations | No | Removed during deduplication |
The count that appears next to the report is therefore a count of safe external destinations, not a count of every clickable object Word can show. A visible URL written as plain text, a citation field, or a clickable image that is not a hyperlink relationship will not appear in the list. The report also does not preserve the anchor text, the page location, or the surrounding paragraph; those details stay in the original .docx.
Privacy, Limits, and a Safer Review Workflow
Privacy is built into the way the extractor works. The browser reads the .docx package locally, opens the relationship XML file, and renders the results in the same tab. No document text, comments, pages, or macros are sent to a remote service, and there is no account, upload, or signup step. The extractor also does not follow any link, so it will not request a website, resolve a redirect, inspect a mail recipient, or decide whether a destination is safe to visit. Each entry in the report is a document-supplied string that still needs normal human review.
There are a few hard limits worth knowing before you start. Only modern .docx files are supported. Legacy .doc files, encrypted packages, and ZIP64 directories with oversized relationship XML are rejected before any result is shown. Internally anchored links, malformed archives, and duplicate targets are excluded by design because treating them as external URLs would be misleading. The tool does not attempt to preserve anchor text, page context, or formatting, and it is not a general document parser; it is a focused relationship-file reader with a stated scope.
For a safer review workflow, keep the original .docx next to the downloaded TXT report, compare the URL list with the visible sources in the document, and then check important URLs with an approved separate process such as a manual visit, a domain reputation tool, or your organization's URL checker. That way the extractor serves as a precise, reproducible starting point while the final judgment about each destination stays with the reviewer. The relationship file format the tool reads is part of the published ECMA-376 Office Open XML standard and is documented in the Microsoft Open XML reference, which describes hyperlink relationships and TargetMode External exactly as the extractor uses them.