Convert PDF to images

Your files and your privacy

The PDF is never uploaded. It is parsed and drawn to a canvas inside this browser tab, and the images are zipped on your own device. The page’s Content-Security-Policy restricts connect-src to this origin, and the PDF rendering engine’s worker file is served from this site rather than a CDN, so no third party sees the document or is contacted while it is open.

What this tool does

Turn each page of a PDF into a picture without uploading the file. Pages are drawn onto a canvas in this browser tab at the resolution you choose, encoded as PNG or JPEG, and packaged into a single ZIP named so the files sort in page order.

Each page is drawn onto a canvas at the scale you choose and encoded as a PNG or a JPEG, one image file per page. PNG keeps glyph edges crisp and suits text; JPEG produces much smaller files and suits pages that are mostly photographic. All the images are packaged into one ZIP, with names padded to three digits so page 2 sorts before page 10 in any file manager. The result is a picture of the page: it looks the same, but the text inside it is no longer text.

How to use it

1. Select Choose a PDF file. The page count appears above the options, which is also the number of images you will get.
2. Pick a format: PNG for documents and screenshots, JPEG when the pages are photographic and file size matters.
3. Pick a resolution — Screen (72 dpi), Good (108 dpi), High (144 dpi) or Very high (216 dpi). Higher settings allocate a much larger canvas per page.
4. Select Run and watch the progress line, which names each page as it is drawn.
5. Select Download to save the ZIP, then read the result note: it says how many images were produced and whether any page had to be rendered smaller.

Example use cases

• Dropping a page of a report into a slide deck or a chat message that will not accept a PDF.
• Producing preview thumbnails of a document for a website or a catalogue listing.
• Sending one page to someone whose phone cannot open PDF attachments reliably.
• Getting a flat picture of a page whose layout breaks when a recipient opens it in a different reader.

Supported input and output

Accepts

• PDF (application/pdf), one file, up to 50 MB
• Unencrypted documents only

Produces

• PNG (image/png) or JPEG (image/jpeg), one file per page
• A ZIP named <original>-images.zip, with entries page-001.png or page-001.jpg

What this tool will not do

• It will not open a password-protected PDF. A document that asks for a password is reported and the job stops.
• Rendering is checked against a memory budget before each canvas is allocated, and a page that would exceed it is drawn smaller than you asked. It is stated in the result rather than hidden, but the images in one ZIP can then differ in scale.
• The output is not searchable and carries no text layer. There is no OCR here, so a screen reader sees nothing but a picture.
• Results are always delivered as a ZIP, even for a one-page document; entries are stored uncompressed, since PNG and JPEG data is already compressed.
• Very high resolution on a long document is slow and memory-hungry on a phone. Cut the pages you need out first if the tab struggles.
• JPEG is encoded at a fixed quality of about 92%; there is no quality slider, and it is not a lossless option for text.

Common mistakes

• Choosing Very high on a long document and assuming every page came out at that scale. Oversized pages are quietly reduced to stay within the memory budget, so check the result note before using the images for print.
• Converting a document that will need to be edited or searched later. Rasterising throws away the text layer, and going back requires OCR that this tool does not do.
• Picking JPEG for pages of small text. The encoder leaves visible fringes around glyph edges; PNG has none and is usually the right choice for documents.
• Expecting the ZIP to be smaller than the PDF. A rasterised page is normally far larger than the vector page it came from, and the archive stores entries without compressing them.
• Treating the images as a flattened copy of the original. Page dimensions become pixel dimensions at the chosen scale, so a mixed-size document produces images of mixed sizes.

Technical notes

Pages are parsed by pdf.js, which is loaded only when this operation runs and whose worker script is bundled from this site rather than fetched from a CDN, with JavaScript evaluation inside the document disabled. Parsing happens in the pdf.js worker; canvas encoding has to stay on the main thread, so the loop yields to the event loop between pages and progress keeps painting. Before each page a viewport is measured at scale 1 and checked against a per-device pixel budget, and the scale is reduced if the raster would exceed it. Each canvas is released immediately after encoding rather than waiting for garbage collection, which matters on iOS where the memory ceiling is per tab. The ZIP is written in the browser with stored entries and CRC-32 checksums.

Frequently asked questions

Are my pages sent to a server to be rendered?

No. Rendering happens on your device, and the PDF engine’s worker file is served from this origin, so no third-party request is made. The page’s Content-Security-Policy limits connect-src to this origin, which you can confirm in your browser’s network panel.

Which resolution should I choose?

Good (108 dpi) suits reading on screen and sharing. Use High or Very high if the image will be printed or zoomed into. On a phone, a long document at Very high is likely to be reduced to fit the memory budget.

Why is one image smaller than the rest?

Because that page was larger than the others, and the raster at your chosen scale would have exceeded this device’s memory budget, so it was rendered at a reduced scale. The result note tells you when that happened.

Can I get a single long image of the whole document?

No. Each page is rendered and saved separately, then zipped. Stitching pages into one tall image would need an even larger canvas than a single page, which is the allocation the memory guard exists to prevent.

Can I convert the images back into a PDF?

Yes, with Convert images to PDF. Be aware that the round trip is lossy: the text became a picture when it was rendered, so the rebuilt PDF cannot be searched or selected.