How to download one or all images from Wikipedia and Commons in full size

Wikipedia articles and Commons category pages show thumbnails, and every image is a file on Wikimedia Commons with a full-size original. For one image, use the download button in Media Viewer (method 1). For every image on a page, PixLasso swaps each Wikimedia thumbnail for its original file and saves them in one ZIP, up to 30 per download in the free version (method 3). It needs no account.

Method 1: Media Viewer or the file page

Click an image to open the Media Viewer and use its download button, which lets you pick the original size. On a file page (File:…) the "Original file" link below the picture does the same.

Method 2: Edit the thumbnail URL

A thumbnail address looks like //thumb.wikimedia.org/wikipedia/commons/thumb/2/25/Name.jpg/250px-Name.jpg?utm_source=…. Remove thumb/, the last part (/250px-Name.jpg) and the ?utm_ query, and change the host to upload.wikimedia.org. The original path on thumb.wikimedia.org redirects to the Commons home page. On 14 November 2026 the original of the Einstein portrait in the English article was 4,530 × 5,607 pixels and 8.9 MB; the 250px thumbnail the page shows was 23 KB.

For a medium size, change the width number instead, but only some widths exist. In our test 20, 40, 60, 120, 250, 330, 500, 960, 1280, 1920 and 3840px returned the image, while 180, 320, 640, 777, 800, 1024, 1200 and 2560px returned HTTP 400. commons.wikimedia.org/wiki/Special:FilePath/Name.jpg?width=777 rounds up to the next width that exists (960px here), and without ?width it opens the original.

Method 3: The whole page into one ZIP

In a test on 23 October 2026, PixLasso found 70 images on the English Wikipedia article "Cat", 53 of them as full-size originals, and all 12 sample files downloaded into a valid ZIP. On the Commons category "Cats" it found 202 images, 186 as originals. There, 3 of 12 sample downloads failed with HTTP 429 (too many requests): Wikimedia limits how fast one address can download. If that happens, wait a minute and download the missing files again.

  1. Add PixLasso to Chrome or Edge and open the page.
  2. Click the PixLasso icon, then Deep scan so it scrolls the page and collects images that load late.
  3. Set Min size to 300px to skip icons and avatars. Originals carry an HD badge.
  4. Click Download. You get one ZIP, numbered in page order, converted to JPG if you choose.

The free version saves up to 30 images per download. See how other sites did on the works-with page.

All of Wikipedia, with images, offline

For the whole encyclopedia rather than one page, use Kiwix. It publishes Wikipedia as single .zim files at download.kiwix.org, which you open in the free Kiwix reader. On 4 November 2026 the newest full English file with images (wikipedia_en_all_maxi_2026-08.zim) was 119 GB, the same without images (nopic) 49 GB, and the intro-only mini 13 GB. Images in these files are resized for reading, so for a print-size original use the methods above.

Copyright

Most Commons files are public domain or under a free license such as CC BY-SA, which usually requires credit to the author and a link to the license. The file page lists both.


PixLasso tested 23 October 2026 in headless Chromium; URL patterns and widths rechecked with curl on 14 November 2026. Written by the developer of PixLasso; the first methods need no extension.