How it works: why your PDF never leaves your device
Choosing a PDF grants this page access to that selected file. The converter reads the bytes into browser memory, then passes them to PDF.js, Mozilla’s JavaScript PDF rendering library. PDF.js interprets the document’s page instructions: text placement, fonts, paths, embedded pictures, page rotation, and other supported graphics. It paints each selected page onto a browser canvas. The browser then encodes those canvas pixels as JPG or PNG. This is rasterization: a page description becomes a rectangular grid of pixels. Mozilla’s PDF.js examples document the document, page, viewport, and canvas rendering steps.
All of that work happens on your device. The PDF engine and its worker are served with this website, using matching version 5.4.149 resources. There is no external PDF worker CDN and no conversion upload endpoint. An opening password is handed to the local PDF engine, not submitted to a server. Downloads are made from browser-created blobs. A multipage ZIP is assembled in browser memory using the same local outputs.
Local processing does not mean that loading the website makes no network requests. Your browser first downloads the page, scripts, worker, and supporting font and decoding resources. Existing site analytics may record ordinary usage events, including file counts and conversion timing, but this converter does not put PDF contents, passwords, or file names in those events. When the tool says it is ready for offline use, you can disconnect and continue in the same tab. Closing or refreshing the tab loses the in-memory document and results; keep the original PDF and download anything you need before leaving.
Choose a DPI for the job
DPI here controls the number of output pixels per inch of the PDF page. PDF.js uses a viewport based on 72 units per inch at its default scale, accounting for the document’s page geometry. The converter scales that viewport by your chosen DPI divided by 72. A larger DPI supplies more pixels for text edges and vector shapes, but it also makes rendering and downloads more demanding. Doubling DPI doubles both dimensions and creates approximately four times as many pixels.
The following values are starting recommendations, not mandatory standards or promises about every document. Check a representative page at the intended viewing or printing size. A large photograph and a dense spreadsheet can need different settings even when their page sizes match.
| Intended use | Suggested starting DPI | What to check |
|---|---|---|
| Screen sharing and quick previews | 96–150 | Labels remain readable at the display size; use 72 for a small thumbnail. |
| Email attachments | 150 | Read the smallest text, then lower JPG quality if the attachment is too large. |
| Printing at the original page size | 300 | Confirm output pixel dimensions and the printer’s requested physical size. |
| Input for a separate OCR tool | 300 | Keep characters distinct; PNG avoids added JPEG artifacts around letters. |
| Posters and enlarged layouts | 150–300 at final size | Calculate pixels from final inches and viewing distance; enlargement needs more source pixels. |
For an A4 example using 8.27 × 11.69 inches, multiply each side by DPI and round to the nearest pixel. At 72 DPI, 8.27 × 72 = 595.44 and 11.69 × 72 = 841.68, or approximately 595 × 842 pixels. At 150 DPI the result is 1240.5 × 1753.5, or 1241 × 1754 pixels. At 300 DPI it is 2481 × 3507, or 2481 × 3507 pixels. These numbers use rounded inch measurements. A real PDF’s page box may differ slightly, and the renderer rounds canvas dimensions down to stay within memory limits.
A DPI setting controls raster dimensions; it is not a guarantee that a JPG viewer or printer will read matching physical-resolution metadata. Browser canvas encoders can write their own resolution metadata. For printing, verify the pixel dimensions and set the physical print size explicitly. A poster printed at twice the width and height needs twice as many pixels along each side to maintain the same effective pixels per inch. Simply selecting 600 DPI cannot invent detail in an already blurry scan.
Custom DPI accepts values from 1 through 600. The converter automatically reduces scale when a page would exceed 100 million pixels or 32,000 pixels along either side, and reports the effective DPI. These are safety ceilings, not a guarantee that every phone can allocate a canvas of that size. One uncompressed RGBA pixel normally takes four bytes before other rendering buffers are considered. Select fewer pages and reduce DPI if your browser reports an allocation error or becomes slow.
JPG or PNG: match the format to the page
JPG is usually a practical choice for photograph-heavy pages, scanned color material, or previews where a smaller download matters. It uses lossy compression: some pixel information is discarded to reduce size. The quality control in this tool runs from 50 to 100, with 90 as the starting value. Try 80–90 for photo-heavy documents, then examine faces, gradients, small captions, and the edges of diagrams. If file size matters more than fine detail, try 70–80 on one page before converting the complete document.
PNG is a better starting point for small text, screenshots, monochrome line drawings, diagrams, and sharp colored shapes. Its lossless compression preserves the pixels produced by the renderer without adding JPEG artifacts. That does not make the whole PDF conversion lossless: fonts and vectors have already become pixels at the chosen DPI. PNG can also be much larger for photographic pages. Choose it for clean edges or transparency, then compare the actual downloaded file size instead of assuming one format is always smaller.
A JPG has no alpha channel. Its background is therefore filled with the color you select, white by default. For PNG, the transparent-background checkbox leaves unpainted parts of the page transparent. A PDF can contain a real white rectangle or a scanned white sheet; enabling transparency will not remove those painted pixels. If a transparent PNG looks black in a particular image viewer, test it over a light background before assuming the conversion lost content.
For additional size control, take the downloaded result to the image compressor. If a form specifies an exact width or height, use the image resizer after conversion. Our guide to JPG quality settings explains compression tradeoffs that also apply to raster images exported from PDFs. Avoid repeatedly opening and re-saving a JPG at lower quality; return to the original PDF when you need a different conversion setting.
Full pages or embedded images: two different results
Render pages creates a picture of each selected page, including supported text and graphics in their visible layout. Use it for receipts, handouts, presentation slides, and pages that must look like the document. PDF text in that picture is no longer selectable text, hyperlinks do not remain clickable, and document structure such as bookmarks does not become an image feature. Retain the PDF if searchability, accessibility structure, or editable text is important.
Extract embedded images looks for supported bitmap painting operations in PDF.js’s page operator list. It can recover a photograph used inside a larger page without rendering the surrounding text. Extraction uses the decoded bitmap’s dimensions, so page DPI is not applied. The result is re-encoded into the chosen format. It is not a byte-identical recovery of an original JPEG attachment, and original camera metadata is not copied.
Not everything that looks like a picture is stored as one bitmap. A chart may be vector paths and text, while a photograph may use a separate soft mask, multiple tiles, or an unusual encoding. Reused bitmap references are exported once per selected page. Unsupported image operations generate warnings; a selection with no extractable images reports that the images cannot be extracted. A partial result names the pages with problems so you can decide whether it is useful. Render pages is the fallback when the visible appearance matters more than isolating individual images.
How this differs from upload-based online converters
The comparison below describes the vendors’ online tools, as stated on their sites, checked 2026-10. It does not describe their desktop products or claim an independent audit of their infrastructure. Product policies and plan boundaries can change. The useful distinction for this workflow is where the document processing happens and whether embedded-image extraction is part of a paid plan.
| Online tool | Where the PDF goes | Published retention or extraction information |
|---|---|---|
| This converter | PDF bytes stay in this browser | No server upload; free page rendering and best-effort bitmap extraction. Results are kept in tab memory. |
| Smallpdf | Its instructions ask users to upload the document | States automatic deletion one hour after processing; full-page conversion is free and image extraction is a Pro feature, with a trial. |
| PDF24 Tools | States that images are created on its cloud servers | States that files are deleted from its server after one hour; offers resolution and image-quality settings. |
| iLovePDF | Its online conversion flow uploads files | Offers page-to-JPG and embedded-image extraction modes. This table makes no retention-time or paid-plan claim for that service. |
An upload-based service can move processing away from a low-memory device, while a browser-local tool uses the computer or phone you already have. Local conversion removes the document-upload step and a remote conversion queue, but performance still depends on page complexity, available memory, and browser capabilities. No universal speed advantage or numerical image-quality ranking is claimed here. Pick a workflow that matches your document handling requirements and inspect the actual result.
Troubleshooting common PDF conversion problems
Blurry text: first increase DPI, then inspect the downloaded file rather than a small thumbnail. Thumbnails deliberately use a low resolution for browsing. If text has small halos or blocks at its edges, try PNG or a higher JPG quality. If the original PDF contains a low-resolution scanned page, increasing output size only enlarges existing pixels. Rescanning the source at a useful resolution is the way to obtain additional real detail.
A download is too large: reduce DPI if you have more pixels than the destination needs. For JPG, reduce quality gradually and check the smallest important text. Remove unneeded pages from the selection. A ZIP groups images for convenience, but already-compressed JPGs and PNGs do not necessarily shrink much more inside it. If an upload form has a strict limit, convert one representative page first and use the compressor or resizer for the final adjustment.
The PDF asks for a password: enter the document’s opening password. An incorrect entry can be retried in the same prompt. If you do not know the password, request an accessible copy from the document owner. This page is not a password recovery service. A PDF can also have restrictions on editing or copying that are different from an opening password; follow the permissions that apply to your document.
A scanned PDF has no editable text: page conversion exports its visible scan as pixels. It does not perform optical character recognition. A scan may include an invisible OCR text layer, but converting the page to JPG does not carry that searchable layer into the image. To feed a separate OCR application, try PNG at 300 DPI while remembering that upscaling cannot repair an out-of-focus source.
A page appears blank or incomplete: try white as the background and disable PNG transparency. White text on a transparent background can appear invisible in a light viewer. Missing or non-embedded fonts, malformed page instructions, unusual transparency groups, and unsupported PDF features can also change rendering. Compare the original in a trusted PDF viewer and re-export with fonts embedded if possible. A file that is truncated or damaged should be downloaded or exported again. Do not treat a blank output as proof that the original page has no content.
The browser slows down or stops: use a smaller page range and reduce DPI. Rendering is serial, and each page canvas is released after encoding, but decoded resources, the PDF itself, and completed image blobs can still consume memory. A ZIP needs the selected outputs to remain available during packaging. The cancel button stops between pages; the current PDF.js render must finish first. Very large pages can fail below the stated safety ceilings on a device with limited memory.
Use it on iPhone, Mac, Windows, or Android
On iPhone or iPad, open this page in your browser, wait for the ready message, and use Choose a PDF to select a file from Files. A document stored only in a cloud drive may need to be downloaded to the device first. Keep the tab in the foreground while converting a large selection, then use the browser’s download interface to save the result. Multiple outputs arrive as a ZIP that can be opened with a compatible file manager. Begin with a few pages if your phone has limited free memory.
On a Mac or Windows PC, use the file picker to select your PDF, review the page thumbnails, and download the result to your usual downloads folder. You do not need to install a PDF printer or configure a server. The file picker chooses a local file; its label does not imply a network upload. If your browser asks where to save a download, choose a location you can find again. Unzip multipage output before attaching an individual page to a form or email.
On Android, use the system document picker and save the JPG, PNG, or ZIP through the browser’s download controls. Actual controls vary by browser and file manager. A small screen uses the same page-range input and options as desktop, so you can select a precise range without checking hundreds of thumbnails. This page supports ordinary keyboard navigation as well: move through labeled controls with Tab, toggle checkboxes with Space, and activate buttons with Enter or Space.
Keep a useful original and prepare images for the next step
A converted image is useful for sharing a page preview, but the original PDF remains the best reference for document structure, signatures, selectable text, and exact page layout. Converting a signed PDF to an image does not create a digitally signed JPG. A visible signature may be drawn into the pixels, while the original signature validation information stays in the PDF. Keep the source when authenticity or later editing matters.
If your source material is a phone photo rather than a PDF, use HEIC to PNG for a lossless raster output or the batch converter for a collection of supported image files. Use this dedicated PDF page whenever you need all document pages or a page ZIP; the batch converter is intended for its supported image formats. For metadata housekeeping after other image workflows, see the EXIF remover and our guide to image metadata and privacy.
Finally, check a downloaded page before sending a whole set. Confirm its orientation, the smallest text, background color, and whether the chosen image format is accepted by the destination. File names retain the source base name and original page number, such as document-page-003.jpg, so a selected subset stays easy to match to the PDF. If the file name itself contains private information, rename the downloaded file before sharing it. Local conversion controls processing location; the way you distribute the resulting images is a separate decision.