Split PDF

Select PDF pages and save them as a new PDF, JPG or PNG images, or TXT. Use OCR for scanned pages. Process documents locally without file uploads.

Guide and frequently asked questions

Extract selected PDF pages, export them as images, or turn their text into a TXT file.

How to use this tool

  1. Choose a PDF and select the pages to keep in the preview editor.
  2. Choose PDF, PNG/JPG ZIP, TXT for embedded text, or TXT (OCR) for scans; choose DPI and OCR language when shown.
  3. Process and download; check page order, image clarity or recognized text.

Example

For an eight-page report, select only pages 2 and 3 and choose PDF: the result is a single two-page PDF, not eight separate files. For pictures choose PNG/JPG ZIP instead. If you need the words from a scanned report, choose TXT (OCR) and its recognition language; ordinary TXT only reads existing embedded text.

Limits and things to check

PDF → PDF, PNG/JPG in ZIP, or TXT. Up to 100 MiB per input, 50 files and 200 MiB per operation; a source PDF may contain up to 500 pages. OCR is limited to 20 selected pages per run. TXT without OCR reads embedded text only. Image export rasterizes pages; requested DPI may be reduced for memory safety.

Frequently asked questions

Why is TXT empty, and when should I use OCR?

TXT reads embedded text and does not recognize a picture of text. Use TXT (OCR) for scans, choose the document language and select no more than 20 pages per run. OCR can lose layout and misread characters; higher requested DPI may be reduced to fit memory limits. If image export exhausts memory, select fewer pages or lower DPI. An empty TXT result does not prove the PDF is empty: it may contain scans or unreadable font mappings. Try OCR on a small selection and verify the words.

Are my files uploaded?

Processing runs in your browser. File contents, names and recognized text are not sent to FileFlow servers or analytics. Loading the site and OCR models still uses the network. Keep your original until you have checked the download.

Practical example and format choices

Choose what you need to preserve

Wanted resultFileFlow outputWhat changes
Only selected document pagesPDFNew document; inspect signatures, forms and bookmarks.
Pictures of whole pagesPNG/JPG ZIPText and graphics become pixels.
Words already selectable in the PDFTXTReading order may differ from the page.
Words visible only in a scanTXT (OCR), or Scan to documentRecognition can make mistakes; original layout is not reconstructed.

Worked example: send pages 2 and 3, then correct the text

You receive a four-page report but a colleague only needs pages 2 and 3. In Split PDF, include those pages, arrange them in the required order and export PDF. Open the download and verify that it contains two pages in that order. A page count alone does not prove the correct pages were selected.

If the colleague needs editable words instead, first try selecting a sentence in the original PDF reader. A usable text layer makes TXT extraction the simpler route. If the words are only an image, choose OCR and the language used on the page.

For a scan that also needs corrections and a Word download, open Scan to document instead. Recognize the selected pages, compare names and numbers against the originals, correct the text and export DOCX. That is a new edited document, not a pixel-perfect copy of the report.

Read the complete guide