Convert arXiv PDF to Markdown
Turn a paper into clean Markdown straight from its link. Equations, tables and the reference list are preserved, so you can read it comfortably, take notes, or feed it to an LLM, without wrestling with the PDF.
From a paper link to Markdown
A research paper is the hardest kind of PDF: dense equations, multi-column layout, tables, figures and a long reference list. Paste the paper's PDF URL and it converts to Markdown with the reading order corrected, the math kept as notation, tables rebuilt, and the references preserved as text. The result is something you can actually search, quote and hand to a model, instead of copy-pasting from a two-column PDF and getting scrambled lines.
Convert a paper in 4 steps
No download, no account. Works from the paper's direct PDF link.
Copy the PDF link
On the paper's page, copy the direct link to its PDF.
Wait for the job
Status goes queued, processing, ready. Equations, tables and references are preserved.
Read or download
Preview the Markdown, copy it into your notes, or download a .md file.
The hard parts handled
Equations kept
Inline and display math is preserved as notation, not flattened into garbled symbols.
Two-column order
Multi-column pages are read in the correct order, so paragraphs do not interleave.
Tables and results
Results tables are rebuilt into real Markdown tables you can read and reuse.
Ready for RAG
Clean Markdown is ideal context for a RAG pipeline or a literature-review assistant.
Building a paper library? Convert many at once with the batch workflow, or drop the Markdown straight into Obsidian or Notion for your reading notes.
What to watch for
Papers are demanding documents, and a converter should be honest about the edges.
Very long papers
A long review or thesis can hit the time budget and return as a partial result, flagged as truncated, so you know it is not the complete document. Splitting it or using a higher tier helps.
Dense or rotated tables
Most results tables convert cleanly, but a very wide table rotated sideways or with deeply merged headers can need a quick visual check.
Hand-written math
Typeset equations come through cleanly. Hand-drawn or photographed math is OCR'd but, as with any tool, is worth a glance for the trickiest expressions.
Build a searchable paper library
One paper is useful; a converted collection is powerful. Because the output is plain Markdown, a reading list turns into something you can search, link and query.
Convert and collect
Run a reading list through the batch workflow to turn a folder of paper PDFs into a folder of Markdown. Drop them into Obsidian or Notion so each paper becomes a linkable note with its equations and tables intact.
Ask across papers
Feed the same Markdown into a RAG pipeline and you can ask questions over your whole library, with the model citing the passages it used. Clean Markdown chunks far better than raw two-column PDF text, so retrieval is more accurate.
Common questions
Can I convert an arXiv paper to Markdown?
Yes. Paste the direct PDF link to the paper, or upload the file. It converts to Markdown with its equations, tables and references intact.
Do equations and references survive?
Yes. Math is kept as notation, tables are rebuilt into Markdown, and the section structure and reference list come through as text rather than scrambled symbols.
Do I need to download the paper first?
No. Convert straight from the PDF URL, so you can go from a paper link to Markdown without saving the file locally.
Is this good for feeding papers to an LLM?
Yes. Clean Markdown is compact and structured, so a paper fits a prompt or a RAG index far better than raw PDF text, with the math and tables intact.
What about very long papers?
Long documents convert. An extremely long one can return a partial result that hit the time budget, flagged as truncated so you know it is not the full paper.
Is it free?
Yes. Convert anonymously in the browser on the free tier (3 slots, 10 MB files, a 15-minute time budget, 1-hour retention). Paid tiers raise every limit.