This is a best-effort extractor for text-based PDFs. Scanned or image-only PDFs contain no text layer and will return nothing — those need OCR, which this browser tool does not do. Complex encodings may produce imperfect characters.
Files are processed locally in your browser and are never uploaded to our servers.
PDF file
Extract text from a PDF
Choose a single PDF. Larger files take a moment because every page stream is decoded in your browser.
Extracted text
Choose a PDF and the extracted text appears here. Line breaks are reconstructed from the text-positioning operators, and a page selection lets you pull out a single page.
How the extractor works — and what it cannot do
The file is read as raw bytes, its objects and stream … endstream pairs are located, and any stream marked /FlateDecode is inflated with the browser’s built-in decompression. The decoded content stream is then scanned for the text-showing operators Tj, TJ, ' and ", with Td, TD, Tm, T* and ET treated as line breaks, and spacing corrections in TJ arrays (numbers of about −100 or lower) turned back into spaces.
Characters are interpreted as WinAnsi/Standard encoding, including the usual escapes (\(, \), \\, \ddd) and hex strings. A font that embeds its own ToUnicode map, or a subset font with a custom encoding, can still produce incorrect characters — this is why the result is described as best-effort. Images, drawings and annotations carry no text here, and text inside images is never recognised.