To get a PDF into Obsidian as an actual note — searchable, linkable, editable — convert it to Markdown and save the .md file into your vault. Embedding the PDF itself (![[report.pdf]]) shows you the pages, but the content stays dark matter: full-text search misses most of it, you can’t [[link]] to a section, graph view ignores it, and highlights live outside your notes.
The conversion takes under a minute and needs no plugin.
The three-step version
- Convert: drop the PDF into the free PDF to Markdown converter — no signup, nothing stored.
- Save: download the
.md(or paste it into a new note) inside your vault folder. - Organize: add frontmatter and links like any other note:
---
source: annual-report-2026.pdf
imported: 2026-07-23
---
# Annual Report 2026
Related: [[Portfolio review]] · [[Q3 planning]]
...
Headings become real Obsidian headings (so [[note#Section]] links work), and tables arrive as Markdown tables that render natively.
What converts well — and what doesn’t
Conversion reads the PDF’s text layer, so digital-native PDFs — papers, reports, statements, ebooks exported to PDF, anything with selectable text — come out clean. Scanned books and photographed pages have no text layer; those need OCR first, which is a different tool class (how to tell the difference, and which tools do OCR).
Two Obsidian-specific tips:
- Big PDFs → split notes. A 300-page book as one note makes Obsidian sluggish. Split the Markdown on
#headings into a folder of chapter notes; each chapter becomes linkable on its own. - Keep the original. Store the source PDF in an
attachments/folder and point to it from the note’s frontmatter — the Markdown is for thinking, the PDF remains the artifact of record.
Batch-importing a folder of PDFs
Migrating a research library? Loop the API over the folder — no key, 100 conversions/day free:
for f in ~/vault/attachments/*.pdf; do
curl -s https://api.marklipi.com/convert -F "file=@$f" \
| jq -r .markdown > "${f%.pdf}.md"
done
Each PDF lands as a sibling .md ready for the vault. (More integration patterns in the Python/Node/curl guide.)
Does this work for Logseq, Joplin, Notion?
Yes — the output is standard Markdown. Logseq and Joplin import it directly; Notion imports Markdown files including tables. Anything that speaks Markdown speaks this.
FAQ
Can Obsidian search inside PDFs without converting?
Core Obsidian search doesn’t index PDF contents; some plugins add partial text search over text-layer PDFs. Conversion is the only route to full first-class treatment — backlinks, headings, graph, edits.
Do the PDF’s images come across?
Text, headings, lists, and tables convert. Embedded images aren’t extracted — keep the original PDF alongside for figures, or screenshot the ones that matter into the note.
Is my PDF uploaded when I convert it?
With marklipi, the file is processed in memory and discarded immediately — never stored, no account involved. For sensitive vaults, that retention answer is the thing to check on any converter you use.
What about annotating PDFs instead?
If your workflow is highlighting rather than rewriting, an annotation plugin against the original PDF may serve better. Converting shines when you want the content itself woven into your notes.