QiLing ToolsⓂ️PDF to Markdown Converter - Headings, Tables & Links Kept
Want to hand a paper, report or manual to ChatGPT, Claude or another LLM, or import it into Obsidian, Notion or a knowledge base? Copying text straight out of a PDF usually leaves broken lines, tables flattened into strings of numbers, and headers and footers mixed in. This tool reads the position, size and font of every piece of text, rejoins broken lines into paragraphs and ranks font sizes into # / ## / ### headings, also picking up bold one-line subheadings. Bullets and numbering become Markdown lists, bold and italic become ** and *, ruled and borderless tables become GFM tables, and hyperlinks on the page become [text](url). Headers and footers that repeat at the top or bottom of most pages, and page numbers such as “Page 3 of 10”, can be removed automatically, and you can add a rule or a <!-- page N --> comment between pages to keep track of sources. If you need images, they can be exported as files, bundled with the .md in a ZIP and referenced in the text. Afterwards you get a live rendered preview, the Markdown source, one-click copy or download, and a character and approximate token count. Everything runs in your browser; the file is never uploaded.
Free · No sign-up · Runs in your browserLast updated:
How to use PDF to Markdown
- 1Upload a text-based PDF (run scanned files through OCR first)
- 2Choose whether to detect headings, tables, links and emphasis, remove headers and footers, export images or add page separators, and which pages to convert
- 3Click Convert to Markdown, check the preview, then copy or download the .md (a ZIP when images are included)
FAQ
- Why is Markdown good for LLMs?
- Markdown expresses headings, lists and tables with very few characters, so a model understands the document’s structure while using fewer tokens than HTML or raw PDF text. Removing headers, footers and page numbers cuts repeated noise; turn on the <!-- page N --> separator when you need to cite page numbers.
- Can it convert scanned PDFs?
- A scan is just images with no extractable text, and the tool tells you when it detects one. Run it through our Scanned PDF OCR tool first to add a text layer, then convert it to Markdown.
- How well are tables and multi-column layouts handled?
- Ruled tables are split along their lines and are the most accurate; borderless tables are detected by column alignment, and merged cells are split. Multi-column pages follow the PDF’s content order, which is the correct reading order for most documents, though complex layouts may need a manual touch.
- Are my files uploaded?
- No. Parsing, conversion and preview all happen locally in your browser, so it is safe for internal documents and contracts.