Import PDFs into Obsidian: convert to Markdown first

Obsidian stores notes as Markdown files on disk. A PDF you drag into the vault is still a binary attachment: you can open it, annotate it with plugins, and embed a page, but you cannot wiki-link into its paragraphs or search its wording the way you search other notes. The reliable path is to convert the PDF to Markdown first, save the .md inside the vault, then link and tag it like any other note. That is how you import PDFs into Obsidian when you care about backlinks, graph views, and local AI. Use PDF to Markdown for PDF-only files. For Word sources, use Word to Markdown. For a mixed pile, start at Convert to Markdown.

Why a PDF in the vault is not Obsidian Markdown

Obsidian’s strength is plain-text notes you can open in any editor. Headings, lists, and [[wiki links]] live in the file. A PDF sits beside that system. Search may find the filename. Full-text search of PDF bodies depends on plugins and indexes that treat the file as a separate object. Your graph does not grow from sentences inside the PDF the way it grows from links you type in Markdown.

People attach PDFs when they need the original layout: a signed contract, a typeset paper, a scanned form. That is fine. Attachment solves archival and visual fidelity. It does not solve note-taking. You still rewrite quotes by hand, or you live inside a PDF annotator that never joins your daily notes.

Markdown conversion flips the job. You pull the text into a .md note, fix the outline once, and weave the content into your vault with links and tags. The PDF can stay in an attachments/ folder if you need the source. The working copy you write against should be Markdown. For why models and local assistants also prefer that text shape, see why LLMs prefer Markdown over PDF.

If your goal is a research vault, treat “import PDF” as “make an editable note,” not “copy the binary.” The rest of this guide follows that definition.

Convert to Obsidian Markdown before you link notes

Conversion is the missing step between a download folder and a useful vault note. You run a PDF-to-Markdown pass, open the result, then move the file into the folder Obsidian watches. After that, every link you add is a normal vault link.

Online conversion fits one-off papers, saved articles, and handouts under a few megabytes. Scripted conversion fits hundreds of files or private corpora you refuse to upload. Microsoft’s MarkItDown library is a common local option; the comparison lives in what MarkItDown is. This article stays on the vault workflow and the online path most readers start with.

Expect imperfect tables and lost figures. Conversion gives you a draft note, not a finished literature review. Budget five minutes of heading cleanup per paper. That cleanup pays for itself the first time you search a phrase that used to live only on page 14 of a PDF.

Name the output for the vault: author-year-short-title.md or a Zettel-style ID your system already uses. Avoid dumping every export as document.md. Stable names make links and Daily Notes references durable when you rename folders later.

Pick PDF to Markdown for PDF-only work

Open PDF to Markdown, upload a text PDF under the size limit, and download the .md. Free accounts get a small daily quota; Pro raises caps if you batch. Prefer digital text PDFs over phone photos of pages. Scans need OCR and more review.

Skim the first screen of the download before you move it into Obsidian. Confirm the title is a # heading and sections use ##. Flat headings become a wall of paragraphs that feel worse in Reading view than the original PDF.

Use Word or the hub when the source is not PDF

Many “PDFs” started as Word. If you still have the .docx, convert with Word to Markdown instead of printing to PDF first. Extra print-to-PDF steps add layout noise and broken columns you then have to repair in Obsidian.

When the folder mixes PDF, Word, and slides, use Convert to Markdown so you do not guess which single-format page to open. Keep PDF to Markdown for true PDFs only. That split keeps each converter’s quirks predictable when you batch a semester of readings.

How to prepare the PDF before conversion

Garbage in stays garbage. Ten minutes on the source saves twenty minutes of vault cleanup.

Prefer a text layer you can select with the cursor. If you cannot select text, you have a scan or an image wrapper. Run OCR in a tool you trust, or accept that names and numbers will need hand fixes. Remove password protection so the converter can read the file. Drop cover art and blank divider pages when they add nothing you will ever link.

Split huge binders. A 300-page handbook can convert, but fixing headings as one note is painful. Cut by chapter or by the topics you already use as Obsidian folders. Keep the original PDF in a read-only attachments path if you need the visual record; write Markdown into your notes folder so re-converts do not overwrite your edits.

Strip secrets you do not want in a cloud upload or in a sync’d vault: internal phone lists, salary tables, live credentials. Redact before conversion when the file is borderline. Privacy details for Toolsy uploads sit later in this post and in is it safe to upload documents online.

Walkthrough: from PDF to a vault note

You need three artifacts at the end: the original PDF (optional archive), a cleaned .md note, and at least one link from a map-of-content or Daily Note into that new file.

Work one representative PDF first. Time the convert-plus-cleanup cycle. Only then decide whether you will batch a whole folder. Scaling a broken outline into fifty notes wastes a weekend.

Keep Obsidian closed or ignore the new file until cleanup finishes if your sync client fights half-written drafts. Many people convert, edit in an external editor, then drop the finished note into the vault. Either order works if you do not leave broken tables sitting in a synced folder mid-edit.

Convert, download, and open the Markdown

Upload the prepared PDF to PDF to Markdown. Wait for the result. Download the .md and open it in a plain editor or Obsidian itself. Fix the title heading. Promote true section titles to ##. Demote running headers that the converter treated as headings. Delete repeated footers and page numbers that landed as body lines.

Check lists: real - or 1. lines beat mashed paragraphs. Fix the first table you care about; leave decorative layout alone. Save under your vault’s naming rule.

Drop the file into the vault and link it

Move or save the .md into the folder Obsidian uses as the vault root (or a subfolder you already index). Create a short alias or title that matches how you will search later. From a map note, add [[your-new-note]]. Add tags your system already uses (#paper, #source/pdf, project tags).

If you keep the PDF, store it under attachments and link it from the Markdown note with a standard file link so readers can open the typeset original. The Markdown note remains the place you quote, summarize, and connect ideas. For deeper RAG-style prep beyond a personal vault, see prepare documents for RAG with Markdown.

Check the note before you rely on search

Do not trust the first download. Obsidian will index whatever you drop in, including OCR junk and footer noise.

Open Outline in Obsidian and walk the headings. If the outline is flat or chaotic, fix headings before you write summaries that point at section anchors. Search for a unique proper noun from the PDF and confirm it appears once, spelled correctly. Search for a number you care about (a date, a dose, a dollar figure) and match it to the source.

Delete lines that are only page chrome. They pollute omnisearch and local AI context if you later feed the vault to a model. When tables look wrong, compare them to the PDF and repair the rows you will cite. Image-heavy pages often arrive as empty holes; add a one-line caption and a link to the PDF page if the figure matters.

Fix outline, callouts, and wiki links

Promote the document title to a single #. Keep section titles as ## so embeds and outline navigation stay stable. Add [[links]] from your literature notes to claims in the new file. Those links are what grow the graph; the PDF attachment never did that on its own.

If you use callouts, wrap warnings or key quotes after conversion. Converters do not invent Obsidian callout syntax for you. A short > [!note] block under the abstract beats pasting the whole abstract into Daily Notes.

When tables and images need a second pass

Wide tables break often. Rebuild the few rows you cite, or keep that data in a separate note. Leave decorative spreadsheet dumps out of the vault if you will never search them.

Figures seldom become useful images in a simple Markdown export. For a dedicated pass on tables and images, see PDF to Markdown with tables and images. Until that cleanup exists, treat figures as “see PDF” and put the takeaway in your own words.

Related jobs: Word, decks, and mixed vaults

A research vault rarely holds PDFs alone. Meeting notes arrive as Word. Talk decks arrive as PPTX. Spreadsheets hold tables you want as Markdown pipes.

Route by source format. Word drafts go through Word to Markdown. Mixed folders go through Convert to Markdown. PDF-only batches stay on PDF to Markdown. That routing avoids double conversion when someone printed a .docx to PDF “for sharing.”

If you also build a local AI knowledge base from the same files, the Obsidian Markdown tree is a strong intermediate. The wave guide building a knowledge base for local AI from documents covers that path. Keep one cleaned Markdown store; point both Obsidian and your retriever at it when you can.

Limits, privacy, and when to keep the PDF only

Conversion will not restore a blurry scan to perfect text. It will not preserve magazine layouts. It will not replace a signed PDF as your legal record. Keep binaries when the layout is the point; convert when the text is the point.

Uploading to any online converter is still an upload. Toolsy processes the file for the response and designs the upload to be discarded shortly afterward (typically within about one hour). Toolsy does not use uploads to train its own models. OCR or AI paths may still send content to external providers for that request. Read the Privacy Policy, plus what happens to files after processing, before you send client or HR documents. For secrets, run MarkItDown locally or use an approved internal pipeline.

Skip online conversion when the file exceeds your plan’s size limit, when DRM blocks reading, or when company policy forbids cloud document tools. In those cases, attach the PDF inside an offline vault and extract quotes by hand for the few paragraphs you need.

Frequently asked questions

How do I import a PDF into Obsidian as Markdown?

Convert the PDF to a .md file first, then save that file inside your vault folder. Open Obsidian and confirm the note appears in the file explorer. Add a [[wiki link]] from a map note so the new file joins your graph. Keep the original PDF in attachments only if you still need the typeset layout.

Can I convert PDF to Obsidian Markdown online?

Yes. Use a PDF-to-Markdown converter, download the .md, and move it into the vault. PDF to Markdown is built for that job on Toolsy. Review headings and tables before you rely on search. Prefer text PDFs over scans for cleaner results.

Is attaching a PDF the same as importing Markdown notes?

No. An attachment keeps the binary; Markdown becomes an editable note with links and tags. Attachments help when you need the original layout or annotations inside a PDF viewer. Markdown helps when you want backlinks, outline navigation, and text search across the vault. Many vaults keep both: PDF for archive, Markdown for thinking.

What is the best way to get PDF to Obsidian Markdown for research?

Start with a digital text PDF, convert it, fix the outline, then link the note from your literature map. Quote claims in your own words and link back to the source note. For local batch jobs without uploads, use a MarkItDown-style CLI and the same cleanup checklist. Online conversion fits samples and one-off papers under the size limit.

Does Obsidian full-text search read PDF contents?

Core search focuses on note files. PDF body search needs plugins or external indexes in most setups. Even then, you still lack wiki links inside PDF paragraphs. Converting to Markdown puts the text in the same format as the rest of your vault. If search of attachments matters, treat it as a bonus, not a substitute for a Markdown note.

How do I handle scanned PDFs for Obsidian?

Run OCR before or during conversion, then proofread names and numbers. Expect broken columns and odd hyphenation on phone photos of pages. Fix the sections you will cite; leave decorative pages out of the vault. When OCR quality is poor, type a short summary note and link the PDF instead of trusting a bad extract.

Can I use Word files instead of PDF for the same vault?

Yes, and you should when the .docx still exists. Word to Markdown keeps headings cleaner than a print-to-PDF round trip in most exports. Use the PDF path only when PDF is the only copy you have. Mixed folders belong on Convert to Markdown.

Will PDF to Markdown keep tables for Obsidian?

Sometimes. Simple tables become Markdown pipes you can edit. Wide or merged cells often break and need a manual repair. Check any table you plan to cite. For a deeper tables-and-images pass, see the sibling guide on PDF to Markdown with tables and images.

Is it safe to upload PDFs before importing them to Obsidian?

Only if the content is allowed under your policy. Toolsy processes uploads for the conversion and designs them to be discarded shortly afterward (typically within about one hour), and does not use uploads to train its own models. That is still cloud processing. For confidential files, convert offline. Details live in the Privacy Policy and the upload-safety checklist post.

Should I convert Markdown back to PDF for Obsidian?

Not for vault notes. Obsidian already works on Markdown. Export to PDF when you need a fixed layout for someone outside the vault. Feeding the vault means keeping .md as the source of truth and treating PDF as an optional share format.

For RAG-style corpora beyond a personal vault, continue with prepare documents for RAG with Markdown. For library vs online conversion, see what MarkItDown is. For model-side reasons to prefer Markdown, see why LLMs prefer Markdown over PDF.