Quick answer
Choose PDF to images in the document tool, select a PDF and run processing to download pdf-pages.zip. The current implementation renders every page at scale 2 and names files page-001.png and onward. It offers no DPI input and cannot restore detail missing from a low-resolution scan.
Identify required pages and delivery requirements
Record the PDF's page count, required page range, accepted image formats, pixel dimensions and file-size limits. If you need only selected pages, follow the PDF page extraction guide to create a working copy first. Conversion currently processes every page in its input document.
Open the original PDF and check text, orientation, stamps and footnotes. Keep the original and necessary access permissions; do not replace an important document with images. If your goal is copying or searching text, use the scanned-PDF OCR and text-layer guide. Image rendering and OCR solve different problems.
Download and inspect the page images
- 1Open the document and PDF tool and select PDF to images.
- 2Choose one PDF and confirm its filename. Begin with a short working copy to check the workflow.
- 3Run processing and wait for pdf-pages.zip. Avoid repeated clicks or closing the tab while it is still working.
- 4Extract the ZIP locally and inspect page-001.png, page-002.png and subsequent files in sequence. These numbers refer to input-document positions, not printed book page numbers.
- 5Keep an extracted folder for review, then upload the required PNG files and inspect the receiving system's result.
The output is a ZIP of separate PNG pages, rather than one continuous image or an editable Word document.
More pixels do not recover missing scan detail
The official PDF.js rendering example explains that viewport scale determines canvas dimensions and incorporates page rotation. KitNelo currently uses a fixed scale of 2 and sizes the canvas from that viewport. This does not establish a single print DPI for every PDF.
Text and vector lines are rendered to pixels at the selected scale. Embedded scans remain limited by their original detail and sharpness. Look at the actual small letters rather than comparing pixel counts alone. If the source characters are blurred, seek a clearer original or rescan. Further reducing file size can also hurt legibility.
Check every page, not merely that the ZIP opens
- Count: the number of PNGs should equal the input PDF's page count. Without an extracted working copy, that includes the entire document.
- Sequence: inspect the first, last and chapter-transition pages and match filenames to the content.
- Orientation and boundaries: check landscape pages, rotated pages, table edges and footers for unexpected clipping.
- Legibility: inspect small text, thin lines, signatures and stamps at the intended viewing size.
- Delivery: inspect the uploaded images at the destination. Verify that it has not resized them excessively or kept only the first file.
Create a small acceptance sample with three pages labeled A, B and C, making the second landscape. Expect three images in that order with the second retaining its landscape appearance. This is a suggested fixture, not a universal processing guarantee.
Understand memory, text-layer and privacy limits
The workflow renders pages and accumulates the ZIP in browser memory. Large pages, complex embedded images and long documents increase demand. If export fails, reduce the page range and retry a copy. There is no universal file-size or page-count threshold promised here, and phones cannot be assumed to handle the same workload as desktops.
PNG does not preserve selectable PDF text, bookmarks, interactive forms or digital-signature validation. An image records appearance rather than backing up those features. Conversion also does not remove sensitive information: visible text, codes and signatures remain visible. Files are processed in the current browser; verify permission before uploading the exported pages elsewhere.
Common questions
Can I select 300 DPI or JPEG in this mode?
The current mode exports PNG at fixed scale 2. It has no user-adjustable DPI or JPEG setting. For a required specification, use a workflow that supports it and check the result again.
Why is the image large but the scanned text still blurry?
Rendering cannot reconstruct detail absent from a low-resolution scan. Compare small text in the source and output, then seek a better source or rescan when necessary.
Can I still search the original PDF text in PNG?
PNG does not carry the PDF's selectable text layer. Keep the PDF for searching. Use an OCR workflow with manual review when you need to recognize scanned text; conversion alone is not OCR.
Further reading & sources
Put the guide into practice
Open the relevant tool, preview the result and check your exported file.


