PDF Extractor & Decompiler Online

Open a PDF and explore its pages, text, images and embedded files. Preview each result and download individual items or everything as a ZIP.

Drop your PDF here
Choose PDF

⚠ Drag-and-drop upload could not load — an ad blocker is most likely blocking it. Please disable your ad blocker and reload the page to upload a file.

This online PDF extractor separates a document into pages, readable text, embedded images and attachments. You can browse the recovered content and download the files you need. Here, “decompile PDF” means inspecting and extracting what the PDF contains; it does not recreate the original Word, InDesign or LaTeX project.

Extract images from PDF

Save pictures embedded in the document, such as images/image-0001.jpg. Reused pictures are listed once and linked from their pages. Suitable JPEGs keep their original encoding; other images are decoded to PNG, including transparency where supported. A full-page preview also includes text and vector drawings; it is different from an extracted picture.

Extract text from PDF

Use this PDF to text workflow to read and download Unicode text for each page, for example pages/page-0001/text.txt. Columns, tables and reading order are best effort. Empty pages do not produce invented text.

View and extract PDF pages

Choose a thumbnail, move between pages and switch Preview, Text and Images in the page viewer. Download page.pdf to save one page as a PDF, or preview.png for a PDF to PNG rendering. Previews use bounded resolution. Page exports preserve page appearance but omit interactive annotations and actions. Range selection, merging and editing are not included.

Extract attachments and embedded files from PDF

Download files actually embedded in a PDF, such as an attachments/0001-data.csv spreadsheet. Original display names, file types and byte sizes appear beside their safe download names. External file references are never fetched. The ZIP includes all completed extracted content.

What does a PDF decompiler recover?

A synthetic three-page document could produce this tree. Each document's output depends on its contents; absent images and attachments are normal.

report.pdf/
  pages/page-0001/preview.png, text.txt, page.pdf
  pages/page-0002/preview.png, text.txt, page.pdf
  pages/page-0003/preview.png, page.pdf
  images/image-0001.png
  attachments/0001-data.csv
  info/document.json, bookmarks.json
  extraction.json

Scanned PDFs, passwords and limitations

Image-only PDFs get previews and recoverable images, but no selectable text. OCR, font export, internal stream inspection and source-project restoration are not included. Password-required or extraction-restricted PDFs receive an explanation; there is no password entry or removal flow.

Default processing limits are 128 MiB input, 100 pages, 4 million pixels per preview, 16 million pixels per embedded image, 200 million decoded pixels total, 500 images, 100 attachments and 64 MiB of expanded attachments. Output is capped at 2,000 files and 256 MiB, with a two-minute processing budget. Worker budgets may shorten processing. Completed files remain available when a later stage fails or reaches a limit; the report identifies omissions.

Frequently asked questions

Can I decompile a PDF?

PDF decompilation here means extraction of stored pages and content. It cannot recover the original editing project.

How do I extract images or save one page?

Upload the PDF, choose a page thumbnail, then open Images or use Download page PDF. The preview PNG contains the whole rendered page; extracted images are the pictures embedded within it.

Why are there no images, attachments or text?

A document may contain only vector drawings, no attached files, or scanned pages without selectable text. Scanned text requires a separate OCR tool. Unsupported content and processing limits are reported explicitly.

Where are uploads processed and when are they deleted?

Files are uploaded to our servers for processing. Use the result's Delete control to remove its artifacts. See our privacy policy for data handling; this tool does not promise browser-only processing or a fixed deletion time.