PDF Text Extractor - LLM Optimized

📄

Drop, paste, or click to browse

PDF files and URLs supported — paste with Ctrl+V / ⌘V

Extensionless or renamed PDFs are fine — the file type is checked by content, not name

All processing happens in your browser — no data leaves your device

No file loaded

Output Format

Format Guide

Markdown: Best for both full-context and RAG chunking. Each page has clear headers with filename and page number.
JSON: Structured format with metadata. Ideal for programmatic processing and custom chunking strategies.
Plain Text: Simple format with page markers. Works with any text processor.

Markdown Optimized for LLM context windows and RAG chunking

JSON Structured data with full metadata

Plain Text Simple format with page markers

Page Range (optional)

Leave empty to extract all pages

Reading Order (for multi-column PDFs)

Auto-detect suits most papers — switch to a fixed count or raw order if columns come out interleaved.

Text Cleanup (applied per page after extraction)

Rejoin words hyphenated across line breaks “apoca-” + “lyptic” → “apocalyptic”. Joins only when the next line starts lowercase, so real hyphenates like “U.S.-based” keep their hyphen. Collapse letter-spaced display type “H E A D I N G” → “HEADING”. Tracked-out titles and running heads otherwise cost a token per letter and never match a search. Unwrap hard-wrapped paragraphs Off by default. Joins a line to the next only when it is full-width for the page, ends mid-sentence, and the next line starts lowercase, so tables, lists and captions stay put. Saves roughly a token per line; spot-check tables afterwards.

Parallel Pages (streaming extraction)

Pages stream in as they're extracted. Higher values pipeline through pdf.js's worker — benefit tapers off (single worker thread); elapsed time is shown so you can compare.

Select a PDF file to begin

💡 URL API

You can also fetch PDFs from URLs:

Paste: Copy a PDF URL, then press Ctrl+V / ⌘V on this page
Drag & drop: Drag a PDF file or URL into the box above
URL param: ?url=https://arxiv.org/pdf/2406.11706
Short form: ?https://arxiv.org/pdf/2406.11706
Cleanup flags: &unwrap=1, &dehyphenate=0, &unspace=0 override the Text Cleanup defaults

📄 PDF Text Extractor

Format Guide

💡 URL API