Extract images from PDF

Why this works

Our extractor pulls every embedded image out of the PDF in its original format and resolution, then bundles them in a downloadable ZIP, named by page.

Most PDFs embed images at their original resolution: when a designer placed a 4 MB photo into an InDesign layout and exported to PDF, the 4 MB image is still inside the PDF. Extract Images pulls each embedded image out as a separate file at its full original resolution — not the downsampled, screen-grabbed version you\'d get from screenshotting the page.

Why this matters versus just screenshotting. A page screenshot captures pixels at screen resolution (typically 72–96 DPI for the standard view, maybe 144–192 DPI on Retina displays). The image in the PDF was almost certainly embedded at much higher resolution — a photo placed in a brochure design is typically 300 DPI to support print quality. If you screenshot the rendered page, you lose 75% of the available image data. Extract Images recovers the original-resolution embeddings.

This is different from PDF to JPG / PDF to PNG. Those tools render whole pages as images (one image file per page, showing the page as it appears on screen). Extract Images recovers the source images that were placed inside those pages — the original logos, photos, diagrams, and graphics as discrete files at their original resolution. For three use cases this distinction is critical: asset recovery (the original photo files have been lost; recover from the PDF), compliance audit (regulator wants every image extracted separately), and design reuse (a client deck contains illustrations you want to extract for reuse).

What extracts. Embedded raster images at their original encoded resolution, in their original format. JPGs embedded as JPGs; PNGs embedded as PNGs; TIFFs embedded as TIFFs. No re-encoding, so the image quality is bit-identical to what the PDF\'s author placed. Vector graphics can be exported as SVG when possible (clean conversion); for vector content that can\'t be cleanly translated to SVG, the tool renders at high resolution (300 DPI) as PNG. Form-field background images, watermarks, and overlay graphics are all detected as separate embedded images.

What doesn\'t extract as images. Text content: text is not an image, so a body of text doesn\'t come out as an image — it\'s in the PDF\'s text layer and would be extracted by Extract Text. Vector logos drawn directly in the PDF: these are part of the page content stream rather than separate image objects, so they aren\'t extracted as images. To extract a vector logo from page content, screenshot the page area and convert to SVG with a vector-tracing tool, or ask the original designer for the logo source file. Patterns and gradients: generated procedurally in the PDF, not embedded as images.

Output options. Original formats keeps each image in whatever format it was embedded as. Useful when you want to preserve transparency (PNGs), photographic compression (JPGs), or specific colour-space encoding (TIFFs). Normalised to PNG: re-encode everything as PNG for consistency — useful when downstream processing expects uniform format. The original-format option is the default because it\'s lossless; normalisation requires a re-encode pass.

Metadata. Extracted images keep any embedded EXIF data — camera model, GPS coordinates if it was a phone photo, copyright tags, embedded thumbnails. For privacy-sensitive workflows (e.g. publishing extracted images publicly), strip EXIF using your image editor before redistribution. For legitimate use, the EXIF often contains useful provenance metadata.

Naming and ordering. Output is a ZIP containing each extracted image, named `image-001-page-1.png`, `image-002-page-1.png`, `image-003-page-3.png`, etc. Numbers are ordered by appearance (top-to-bottom, page-by-page) and the page where each image appears is included in the filename for traceability. A 50-page document with 200 embedded images produces a ZIP with 200 image files clearly named.

How it works