Turn a PDF into Markdown so an LLM reads it for a fraction of the tokens, then edit the result. Your file is converted inside this browser tab and never uploaded.
or drop it anywhere on this page
Already have Markdown? Open a .md file · Write from scratch
Attached to an LLM, each PDF page is rendered to an image and read as text — roughly 3,000–7,800 tokens. The same page as Markdown is about 300–500.
The result opens in a live editor with a side-by-side preview. Fix what the extractor got wrong, then save straight back to the file.
Your file is read by this tab and never sent to a server — not to us, not to anyone.
Headings, bold, lists and links survive the conversion. Running headers, footers and page numbers are stripped, because repeating them on every page is exactly what makes a PDF expensive for a language model to read.
Drop the PDF onto this page. The .md is generated in your browser within about a second, and you can copy it or download it as a file. There is no upload, no account and no queue.
A .md file is a plain text file written in Markdown — a way of marking up structure with ordinary characters, so # Title is a heading and - item is a bullet. It opens in any text editor, and tools like Obsidian, Notion and GitHub render it as formatted text. Because it is plain text, it is also the format language models read most reliably.
Select or drop your PDF above. The Markdown appears in an editor within a second or two, and you can copy it or download it as a .md file. Conversion happens inside this browser tab, so the file is never uploaded.
When you attach a PDF to an AI model, every page is rendered to an image and its text is extracted — roughly 3,000–7,800 tokens per page. The same page as Markdown is about 300–500 tokens, because there is no image and no repeated layout scaffolding. On real documents that is an 85–97% reduction.
Here, yes — because nothing is sent anywhere. The conversion runs in JavaScript inside your own tab, so contracts, invoices and bank statements never reach a server. You can prove it: load this page, disconnect from the internet, and convert a file. It still works.
Not in the browser. A scanned PDF is a stack of page images with no text layer, so there is nothing to extract — it needs OCR first. iLoveMD detects this and tells you, rather than handing back an empty file.
Open a .md file or paste Markdown into the editor, then choose Download PDF. Your browser prints the formatted preview to PDF, so again the file stays on your machine.
Convert it to Markdown first, then paste the text. Attaching the PDF makes the model pay for a rendered image of every page on top of the text. Markdown keeps the headings, lists and tables the model needs to understand structure, and drops the layout it doesn't.