How to Convert PDF to Markdown for ChatGPT, Claude, and Obsidian

TABLE OF CONTENTS

PDF text often arrives in the wrong reading order when you copy and paste it, while headings and lists lose their structure. Converting the file to Markdown gives you editable plain text that is easier to reuse in an AI chat or a notes app. Check whether the PDF contains selectable text, then follow the four steps below.

Check Your PDF Before Converting It

Try to select one sentence in your PDF viewer. If you can highlight individual words, the file is probably text-based and ready to convert. If dragging selects an entire page as one image, the PDF is scanned and needs OCR first.

PDF type	Quick test	What to do
Text-based PDF	Individual words are selectable	Convert directly to Markdown
Scanned PDF	The page behaves like one image	Run OCR, review the extracted text, then convert
Mixed PDF	Some pages have selectable text and others do not	OCR the image-only pages before combining the content

OpenL’s PDF to Markdown converter does not process scanned or image-based PDFs. For those files, choose an OCR method from our comparison of the best OCR tools for scanned PDFs before continuing.

Convert PDF to Markdown in 4 Steps

Open the PDF to Markdown converter. Go to openl.io/pdf-to-markdown. The free tool allows up to 30 conversions per day without an account.
Upload your text-based PDF. Click Upload PDF File and select the document, or drag it into the upload area. Do not upload the file if it contains information you are not permitted to send to an online service.
Convert the document. Start the conversion and wait for the Markdown result. OpenL says it preserves the logical structure of headings, paragraphs, bullet and numbered lists, tables, and hyperlinks.
Copy or download the Markdown. Copy the result when you only need it for one prompt, or download the .md file when you want to edit, archive, or import it elsewhere.

OpenL states that uploads use encrypted connections and that PDF files are deleted from its servers after conversion. You should still follow your employer’s or school’s data-handling rules before uploading confidential material.

Check the Markdown Before Using It

Conversion removes much of the manual formatting work, but PDF layout can still be ambiguous. A PDF stores content by position on a page; Markdown stores a reading order and text structure. Review the parts where those models do not map cleanly.

Element	What to check	Fast repair
Headings	Levels follow the document hierarchy	Change the number of `#` characters
Paragraphs	Multi-column text reads in the correct order	Move misplaced paragraphs
Lists	Items are not merged or split	Add `-` or renumber the affected lines
Tables	Rows and columns remain aligned	Repair pipe characters or use a plain list
Links	Link text points to the intended URL	Correct the destination inside `( )`
Footnotes	References still match their notes	Rename duplicated or missing identifiers
Page headers	Repeated titles and page numbers are absent	Delete repeated lines

Search the output for unusual character substitutions, broken words at line endings, and repeated headers. For documents containing equations, dense tables, or sidebars, compare each affected section with the original PDF rather than trusting a clean-looking preview.

Use the Markdown in ChatGPT

ChatGPT supports common text and document file extensions, although availability and upload limits depend on the plan and account settings. Use either of these workflows:

Paste a short Markdown document into the prompt. Put your instruction before the content and wrap the Markdown in a fenced block if you need to distinguish it from the instruction.
Upload the .md file when your ChatGPT account accepts it. If the extension is rejected, save a copy as .txt without changing the contents.

Ask for an output that uses the existing structure instead of a vague request such as “analyze this.” For example:

Summarize this document by its existing headings. Keep every date and number exactly as written, and list any sections that appear incomplete.

Converting to Markdown is most useful when you need reusable text and visible hierarchy. It is not automatically better than uploading the original PDF when charts, diagrams, or page layout carry essential meaning. OpenAI notes that most ChatGPT plans use text-based retrieval for document files and discard embedded images; ChatGPT Enterprise offers visual retrieval for PDFs.

Use the Markdown in Claude

Anthropic’s current upload guide lists TXT, PDF, DOCX, HTML, CSV, ODT, RTF, EPUB, JSON, and XLSX, but it does not list .md. The reliable options are therefore:

Paste the Markdown directly into Claude for a short or medium document.
Save a .txt copy and upload it when the content is too long to paste comfortably.

Keep the Markdown syntax unchanged inside the text file. Claude can still read the headings, lists, tables, and links because the underlying content is plain text.

For a review task, use a specific instruction such as:

Review this Markdown against the source document. Flag broken tables, abrupt paragraph transitions, missing footnotes, and headings that appear out of order. Do not rewrite the content yet.

Claude can also analyze PDFs directly, including visual elements in PDFs of 100 pages or fewer. Use the original PDF when visual context matters; use Markdown when you want to edit, reorganize, version, or reuse the extracted text.

Add the Markdown to Obsidian

Obsidian uses Markdown for notes, so no paste-and-reformat step is necessary.

Rename the downloaded file clearly. Use a short filename such as research-paper-notes.md instead of the converter’s default name.
Move the file into your Obsidian vault. Put it in the folder where you want the note to appear. Obsidian will detect the .md file.
Open the note and repair internal links. Standard web links work as written. To connect the note to another vault note, use an Obsidian wikilink such as [[Project Index]] or a Markdown link to another .md file.
Add properties only if you use them. Metadata such as tags, source URL, author, and publication date can make imported documents easier to find, but it is not required for the note to work.

If you plan to translate the converted document while keeping its markup intact, follow our guide to translating Markdown without breaking its structure.

Fix Common Conversion Problems

The paragraphs are in the wrong order

This usually happens with multi-column layouts, sidebars, or floating captions. Move the blocks into reading order manually and compare the result with the original page. If layout preservation matters more than reusable text, use a workflow designed to translate a PDF without losing formatting instead of flattening it into Markdown.

The table is unreadable

Markdown tables need a consistent number of columns in every row. Remove decorative cells, merge split lines, and check whether a list communicates the information more clearly. Do not keep a broken table merely because the source used one.

Text is missing from a scanned page

Stop and run OCR on the original PDF. Converting an image-only page cannot recover text that the converter cannot see. Review names, dates, totals, formulas, and identifiers after OCR because a single recognition error can change their meaning.

Images disappeared

Markdown can reference images, but a text conversion does not guarantee that embedded PDF images will become separate usable files. Keep the original PDF beside the Markdown when figures, diagrams, signatures, or page design matter.

PDF to Markdown Quality Checklist

Headings follow a logical hierarchy.
Paragraphs from columns and sidebars are in reading order.
Lists and tables contain every item from the source.
Links, footnotes, dates, and numbers match the PDF.
Repeated page headers and page numbers are removed.
Images or charts needed for interpretation remain available separately.
Sensitive content is handled according to the applicable policy.

FAQ

Is OpenL’s PDF to Markdown converter free?

Yes. OpenL states that you can convert up to 30 PDF files per day without creating an account. Higher-volume use requires a paid plan.

Can it convert a scanned PDF?

No. The converter requires a text-based PDF with selectable text. Run OCR first, correct the extracted text, and then convert it to Markdown.

Why not upload the original PDF directly to ChatGPT or Claude?

Both tools accept PDFs. Keep the PDF when you need charts, images, or page layout. Convert to Markdown when you need clean editable text, explicit heading structure, smaller reusable notes, or a file that can live in a code repository or Obsidian vault.

Does Markdown preserve the original page layout?

No. Markdown preserves logical text structure, not fixed page geometry. Headings, lists, links, paragraphs, and simple tables can transfer well, but columns, floating elements, and precise visual placement do not.

Sources

OpenL PDF to Markdown Converter — workflow, free allowance, preserved elements, scanned-PDF limitation, and file handling
OpenAI File Uploads FAQ — supported document workflows, limits, and PDF image handling
Anthropic: Upload Files to Claude — supported formats, file limits, and PDF processing behavior
Obsidian Basic Formatting Syntax — headings, lists, links, and other supported Markdown syntax
Obsidian: How Obsidian Stores Data — Markdown files, vault folders, and automatic refresh after external file changes