The DOCX to Markdown Converter does not open PDF files itself, so to bring a PDF into Obsidian as editable Markdown you re-save the PDF as a Word .docx and then convert that file locally in your browser — no upload, no Pandoc, no command line. The tool reads the Office Open XML package inside the .docx using a JavaScript ZIP reader, parses the main document with the browser's built-in XML parser, and emits a Markdown file that follows your real heading styles and table structure. Headings 1 through 6 in Word become # through ###### in the output, ordinary paragraphs stay ordinary text, and a simple rectangular table is rewritten as a Markdown pipe table with the first row treated as the header. Anything Word can express but Markdown cannot — page numbers, columns, shapes, images, comments, tracked changes, merged cells, custom fonts — is intentionally dropped or downgraded to plain text instead of being mangled into fake syntax. Because every step runs in the current tab, neither the original PDF nor the intermediate .docx ever leaves your device, which matters when the document is a private journal, a draft contract, or anything under NDA. The result is a .md file you can drop into your Obsidian vault and edit like any other note.
Why PDFs sit dead in an Obsidian vault
Obsidian stores every note as a plain Markdown file inside a folder you choose, and the features the app is known for — backlinks, the graph view, full-text search, tag autocomplete, the Dataview plugin, and a long list of community plugins — all operate on the text inside those .md files. A PDF you drop into the same folder is treated as a binary attachment. Obsidian's bundled reader shows the pages, but the words on them never enter the search index, the graph cannot see links between the PDF and your other notes, and you cannot copy a single sentence out of the attachment without switching to a separate PDF reader. That is why people who want a PDF to actually live inside their knowledge base convert it first. The challenge is that the obvious converters either upload your file to a remote server, demand a working Pandoc install plus a LaTeX toolchain, or return Markdown full of broken tables and missing headings.
The two-step bridge: PDF to DOCX to Markdown
The most reliable way to recover a clean Markdown draft from a PDF without giving up privacy is a short two-step bridge. In the first step you open the PDF in Microsoft Word, LibreOffice Writer, or Google Docs and use Save As to produce a .docx file. Word and Writer treat the PDF as a layout to recreate, which usually preserves the text order and gives you a chance to fix obvious OCR errors before they ever enter your vault. In the second step you point the DOCX to Markdown Converter at that .docx and let it read the Word document structure directly, instead of trying to guess at fonts and margins from the PDF. The result is a Markdown file that follows the headings and tables you actually wrote in Word, not what a PDF parser thought they might be. For readers who want a deeper walk-through of the PDF-to-DOCX half, the guide on converting PDF content to Markdown via DOCX covers the same workflow in more detail and lists the free tools that handle that first hop on every operating system.
Prepare the DOCX so headings survive
The converter only respects structural intent, not visual formatting, so a paragraph that was made bold and 18-point by hand will arrive in the Markdown file as an ordinary paragraph. The tool never guesses heading levels from font size or bold text, because guessing would invent structure the source never had. Before you convert, open the .docx and apply Word's built-in Heading 1, Heading 2, and Heading 3 paragraph styles to the titles and section headers you want to see as #, ##, and ### in Obsidian. The Styles pane (Alt+Ctrl+Shift+S in Word, or View > Styles in LibreOffice) makes this a quick sweep down the document. Apply Body Text or Normal to anything that is just a paragraph. Check tables too: collapse merged cells into a single header row where you can, flatten nested tables into separate side-by-side tables, and remove any decorative borders that exist only for the printed page. Complex list numbering in Word cannot be represented faithfully in basic Markdown and falls back to readable tab-separated text instead of an invalid or misleading table.
Convert the DOCX with the DOCX to Markdown Converter
Now the conversion itself. The whole job runs in your browser tab, so close any other heavy pages first if your device is short on memory.
- Open the DOCX to Markdown Converter in your browser.
- Click the file picker and choose the prepared .docx from your device.
- Wait a moment while the DOCX package is checked for size, ZIP directory shape, entry count, and declared expanded size, and the XML parser walks the main document.
- Read the read-only Markdown preview that appears, scrolling through it to confirm that your headings came through as #, ##, ### and your simple tables are clean pipe tables.
- If anything looks off — a table fell back to tab-separated text because it had merged cells, or a heading is missing because it was never styled — fix the source .docx and re-run the conversion rather than trying to repair the Markdown by hand.
- Click the download button to save the .md file to your device.
- Open the .md in any plain text editor or directly inside Obsidian and do a short editorial pass before treating it as a finished note.
The download is a .md file ready to be moved into your vault.
What the converter keeps and what it drops
It helps to know in advance what survives the trip and what does not, so the table below lists the behavior the converter commits to under the Office Open XML structure documented at the Microsoft Learn reference for WordprocessingML documents.
| Word feature | What the Markdown file contains | Why |
|---|---|---|
| Heading 1–6 paragraph styles | # through ###### heading lines | Explicit structure maps to explicit Markdown syntax |
| Bold or large-font text without a heading style | Plain paragraph | The tool does not guess headings from appearance |
| Simple rectangular tables | Pipe table, first row as header | Basic Markdown only supports one grid shape |
| Merged, nested, or uneven tables | Tab-separated plain text | False tables would mislead readers and Obsidian plugins |
| Ordinary paragraphs | Text with line breaks and tabs preserved | Visible whitespace survives instead of being deleted |
| Images, shapes, charts, text boxes | Not included | Markdown has no native equivalent; use a dedicated extractor |
| Comments, tracked changes, fields, citations | Not included | These are review metadata, not document content |
| Custom paragraph or character styles | Treated as ordinary text | Guessing would invent structure the source never had |
If your goal is to capture the readable text without any Markdown syntax at all, the DOCX to Text Converter does that in the same local, no-upload style. If you specifically need the hyperlinks from the Word file, the Word Hyperlink Extractor pulls them out cleanly. Keeping these tasks in separate tools is intentional — a single converter that claims to do everything would quietly drop fidelity somewhere you cannot see, and a small editorial pass after conversion is part of the contract.
Drop the .md into Obsidian and polish it
Once the .md file is on your device, move or copy it into the folder Obsidian watches as your vault. The app indexes it on the next refresh and the title becomes the note's display name. From here the editing is ordinary Markdown: add [[double brackets]] to link the imported note to existing notes, drop #tags for grouping, and use [[note name|alias]] if the imported heading is too long for an inline link. If the conversion fell back to tab-separated text for a tricky table, rebuild it as a pipe table by hand — Markdown tables are simple once the data is on screen. The readability score checker can give you a quick sense of how dense the imported note is, and the word counter is handy when you want to know whether the imported chunk is worth breaking into smaller notes so your graph view stays useful. Save the note in Obsidian and it is now a first-class citizen of your vault, searchable and linkable in the same way as anything you typed yourself.