Extract images from a PDF — pulling the source images out vs rasterizing the whole page
"Extract images" sounds like one operation but is actually two — and picking the wrong one wastes a lot of time. Our Extract Images tool pulls the SOURCE images embedded inside the PDF at their original resolution. Our PDF-to-PNG tool, on the other hand, renders every page as a fresh image. The two produce different output for the same input. Here is when each is right, what a PDF's image inventory actually looks like under the hood, and the three cases where extraction either returns less than you expected or more.
Source extraction vs page rasterization
When a PDF contains a photograph, the image data is embedded as a binary stream inside the file — typically as JPEG or PNG bytes. Extract Images parses the PDF, finds every such embedded image stream, and decodes it to PNG at its native resolution. The output is exactly what the author put into the PDF, no more, no less.
Page rasterization — what PDF-to-PNG does — renders the page exactly as you see it in a viewer, at whatever resolution you pick. The output is one image per page, combining all the text, vectors, and embedded images into a single rasterized bitmap. If a page contains three small photos and a lot of text, page rasterization gives you one page-sized image with everything together; source extraction gives you three small standalone photos.
The right tool depends on what you need next: source extraction for the photos themselves (to use elsewhere, archive separately, run through an image pipeline), page rasterization for an image of how the page looks (to embed somewhere, archive visually, generate thumbnails).
When source extraction is the right answer
Specific cases where the embedded-image format is exactly what you want:
- Recovering photos from a phone scan that went to PDF. If someone sent you a multi-page PDF of photographs taken with their phone, the original JPEGs are still inside the PDF at their full camera resolution. Source extraction recovers them. Page rasterization would re-render them, losing detail.
- Pulling product photos out of a catalog PDF. Marketing teams often produce a single PDF that contains every product image at print resolution. Source extraction pulls each photo as its own file, ready for the website CMS.
- Auditing what images a vendor PDF contains. Sometimes you receive a PDF and want to inventory the embedded media (e.g. checking that high-res images are present for a print-ready document). Source extraction surfaces every embedded image with its dimensions and format.
- Extracting medical or technical diagrams. Scientific papers and medical reports sometimes embed high-resolution diagrams as standalone images. Source extraction preserves the diagram resolution for slide reuse or re-publication.
- Recovering original assets after losing the source files. If you misplaced the originals but still have the production PDF, source extraction is the recovery path.
Three cases where the output surprises you
Friction points worth knowing before you click extract:
- PDFs that are mostly text return zero images. If the PDF was generated digitally from Word or Google Docs without any embedded photos or screenshots, there are literally no embedded images to extract. The output is empty. This is correct behavior — you wanted page-rasterized images, not source images. Use PDF-to-PNG instead.
- A scanned PDF returns ONE image per page, regardless of visible content. Scanned PDFs are usually image-only — each page is a single large rasterized image of the original paper. Extract Images returns those page-sized rasters. If you wanted the diagrams or photos inside a page extracted as separate images, you cannot do that from a scan; the scan only has one image per page by construction. Run OCR first if you need text-level access.
- Vector graphics aren't returned. PDF supports vector graphics as path objects, not as images. Logos, line drawings, and charts created in tools like Adobe Illustrator are typically vector and not part of the image inventory. They render perfectly in PDF viewers but cannot be "extracted" as raster images. If you need them as raster, use PDF-to-PNG (page rasterization) and then crop.
What the output looks like
Each extracted image becomes a separate PNG file. The PNG is decoded from the source embedded stream — JPEGs are decoded to PNG so the output is uniformly PNG (this is by design; raw JPEG sometimes carries embedded color profiles or progressive-encoding flags that confuse downstream tools). File size of each output matches the input image's pixel count and content type.
Filenames carry the source page number so you can see the order at a glance. If a page has multiple images, they are suffixed with an index (page-3-image-1.png, page-3-image-2.png, etc.). The default UI lets you download each image individually; the Download All button bundles every extracted image into a single .zip.
Limits and compatibility
On the free web tool, Extract Images handles PDFs up to 100 MB with no page-count cap and no image-count cap. PDFium runs in WebAssembly in your browser; nothing is uploaded. Rare codecs (JBIG2 under encryption, certain JPX/JPEG2000 variants) can fail to decode — when that happens we log and skip the offending image rather than failing the whole extraction, so you still get every image that did decode.
Common pairings: Pair with Extract Pages to focus on a subset of pages first if you only want images from a specific section. Pair with PDF Inspector to preview the image inventory (count + resolution per page) before clicking extract.