Kiwix: Offline Browser

Official site: https://www.kiwix.org 

• About this app
• Usage
• Search usage
• Image download
• Privacy policy
• Technical information
  • ZIM archive format
  • Downloading and storing large archives
  • JQuery and ServiceWorker modes
• Feedback/helping/contributing
• Credits
• Licence information
• Other platforms/versions
• Download ZIM archives (content):
  • Library of archives 
  • Index of all files 

About this app

If your Internet access is expensive, slow, unreliable, controlled or censored, you can still gain free offline access to many valuable resources. Carry the world's largest Encyclopaedia around in your pocket, in any language, or a worldwide Travel Guide when you are on the road!

To use this app, you will need to download, completely free, one or more archives (called ZIM archives) from the Kiwix website and store them on your device or a thumb drive / external drive. You will then be able to access your chosen content totally offline.

Archives are available with content in many different languages from sources such as Wikipedia, Wiktionary, WikiMed, Wikivoyage, Project Gutenberg, Stack Exchange, TED Talks and many others. There are also "themed" archives with selected topics of interest such as medicine, movies, history, geography, maths, physics, etc.

↑ Back to Contents

Usage

Step 1: Choose your content

Go to https://library.kiwix.org  (easy-to-use graphical interface, archive previews) or https://download.kiwix.org/zim/  (traditional file list), and select the archives that interest you in your favourite language(s). If you don't have Internet access, you can also access/share content from a USB stick.

For a quick test, we recommend you start with a small archive containing all the Wikipedia articles about climate change  in English.

Wikipedia archives with "mini" in the filename are the smallest, since they only contain the introduction of each article; "nopic" means no pictures; "maxi" means everything except video and audio. The largest files without any of these qualifiers are complete. Note that you can only see the filenames easily in the traditional file list .

Before downloading non-Wiki ZIM types, check if your browser supports ZIM archives with dynamic content: look at the bottom of the Configuration page, in the API panel, and if you see "ServiceWorker API available and registered" in green, then the app can read all ZIM types (except Zimit: see ZIM archive format for more details). Otherwise you should mostly stick to MediaWiki (Wikipedia, Wikivoyage, Wiktionary...) and Stack Exchange archives (see modes below for an explanation).

Step 2: Download or copy the content onto your device

Download your chosen content and/or copy it into a dedicated ZIM folder on your device's storage. Check the size before you download archives as many file sizes are very large (several Gigabytes). Themed versions such as WikiMed, or versions without images, are much smaller than full Wikipedia archives and can be readily downloaded in your browser.

If you don't have enough internal storage, consider using a modern microSD card, an external hard drive, or thumb drive. For larger archives (greater than 2GB) on older devices, it is possible that an archive might not be supported if your storage is formatted using a FAT filesystem (NB exFAT is fine). See Downloading large archives below for a solution.

Step 3: Drag and drop the file into the app

You can drag and drop the archive file into the the open app (if you have a split file, be sure to drag all parts), or else you can go to the "Configure" menu and select your ZIM file.

Step 4: Enjoy your offline content!

↑ Back to Contents

Title and full-text search usage

There are three types of search:

Title search matches the start of an article title, so if you search for 'France' you will get 'France (country)', 'France (disambiguation)', 'France (film)', etc. This kind of search tries to be case-insensitive, but the number of case variants tried is necessarily limited. If not enough results are returned, you can increase the maximum number of search results using the slider in Configuration.

Alphabetical search: If you type a letter of the alphabet in the search box (upper case may be most useful), it will show an alphabetical list of articles starting with that letter. This is useful for ZIM archives that have descriptive rather than semantic titles (e.g. TED Talks). You can also display a list of all titles (up to the limit set in Configuration) simply by clicking the search button without typing anything.

Full-text search: This type of search is only availalbe in relatively recent desktop browsers, and if the archive contains a full-text index. If it is supported, it will be used automatically, but it is slower than title search, and you may notice a delay before results are added to the title-search results. Additionally, full-text search cannot be used if your archive is split into various chunks. To see whether it is available, look in the API panel at the bottom of the Configuration page after you have loaded a ZIM. If it says "Xapian [fulltext]" then it should be automatically enabled for that ZIM.

↑ Back to Contents

Image Download

Unfortunately it is not usually possible to download and save an image from a ZIM article using the context menu (after right-clicking or long-pressing the image). One workaround is to choose "Copy Image" in the context menu, then paste it in any suitable software (Paint, Gimp etc, or even a Word processor), and then save it from there.

↑ Back to Contents

Privacy policy

Short version:

Kiwix JS works offline, and does not collect or record any of your personal data. It only remembers your browsing history for the duration of a session (for the purpose of returning to previously viewed pages). This history is lost on exiting the app and is not recorded in any way.

Longer version:

This application only reads the archive files that you explicitly select on your device together with files in its own package. Some ZIM archives contain active content (scripts) which may, in rare circumstances, attempt to contact external servers for incidental files such as fonts. We have a Content Security Policy which attempts to prevent this, but it is not possible to guarantee that all such requests will be blocked, for example if a ZIM were to contain intentionally malicious code that can break the sandbox. Please note that scripts will only run if you enable ServiceWorker Mode (the default) in Configuration. Nevertheless, if you believe your Internet access is insecure, or is being observed or censored, we recommend that you completely shut down your Internet (Data or WiFi) access before using the application.

Additionally, if you obtained this app from a third-party site (including a browser extension store), then the site operator may track your usage of the app (e.g. download, install, uninstall, date and number of sessions) for the purpose of providing anonymous, aggregate usage statistics to developers. If this concerns you, you should check the relevant site's Privacy Policy for further information.

↑ Back to Contents

Technical information

ZIM archive format

Offline archives use the OpenZIM format , but note that this app has only been tested extensively on MediaWiki (Wikipedia, Wikivoyage, etc.), Gutenberg, Stackexchage and TED Talks archives. Video content is supported if your device, browser or OS understands the format. Some ZIMs use a proprietary dynamic UI which is only supported in ServiceWorker Mode (see below).

This application is not currently compatible with archives that are created by Zimit (https://youzim.it ) using the Web Archive (WARC) format. We have plans to enable this format, but there are considerable technical challenges to overcome in order to display these archives with the fidelity that you expect from Kiwix -- see issue #1009 . However, there is experimental support for Zimit archives in our Progressive Web App (PWA) available from https://pwa.kiwix.org .

Downloading and storing large archives

Full versions of Wikipedia can be enormous: French, German and Spanish Wikipedia with images are each >25GB; full English Wikipedia with images is >90GB. For these very large archives you can find convenient BitTorrent links provided by library.kiwix.org , which may be more reliable than attempting to use your browser. You can use an open-source BitTorrent client such as QBittorrent  with these links (install the software before clicking on the torrent link).

If you need to store a large ZIM archive on an older filesystem formatted as FAT16 or FAT32, you need to be aware of the file size limits of those systems (FAT16 < 2GiB; FAT32 < 4GiB). Most modern microSD cards, thumb drives or hard drives are formatted as exFAT or another modern FS such as NTFS, which do not have this issue. If your ZIM archive is larger than the FS limit, it is possible to split the archive into several 2GiB-1 or 4GiB-1 files (or smaller). You will need to give a file extension to each chunk in the right order following this pattern: *.zimaa, *.zimab, *.zimac, ..., etc.). When you pick this archive in the app, be sure to pick all the chunks, or drag-and-drop them all into the app.

For some information on how to split archves, see Download a ZIM archive .

↑ Back to Contents

ServiceWorker and JQuery modes

Depending on your browser or framework, this app may be capable of running in different modes, which we call "ServiceWorker Mode" and "JQuery Mode". There is a toggle under Compatibility Settings in Configuration that allows you to select between these. Here is a technical explanation of what these modes do:

• ServiceWorker Mode: This is the default mode. As its name implies, it requires that the browser or framework be capable of installing a Service Worker, which is usually the case in modern browsers. It works by intercepting the browser or framework's Fetch calls (network requests) and supplying the requested content from the ZIM. In this mode, the content is read and supplied as-is from the archive to the browser. Dynamic content (e.g. JavaScript) and proprietary UIs are fully supported in this mode. However, Zimit archives are not currently supported due to a conflict between the Kiwix JS Service Worker and the WARC Service Worker (but you can use the PWA for experimental support, see above). This mode can feel initially a little slower than JQuery mode until commonly used assets are cached, but it soon equals JQuery mode in speed, at least in modern browsers. However, older browsers such as IE11 are incompatible with this mode, and the app must be running in a secure context (https:, localhost, or certain browser extensions). While this mode is not natively supported in Mozilla (Firefox) browser extensions, we provide a functional workaround by re-launching the extension as a Progressive Web App (PWA). Note that this mode cannot run with the file: protocol (but only IE11 and old Edge allow the app to run by launching index.html from the file system).
• ServiceWorkerLocal Mode: This is only available in Chromium extensions. It is the same as above, but due to very tight security restrictions in these extensions, no inline JavaScript or eval statements can run. Unfortunately, modern ZIMs may have such dynamic content, and so functionality may be restricted in this mode. Nevertheless, it is more secure: the content is sandboxed and no one-time server access is required to enable the Service Worker. You are protected from inline code execution at the cost of loss of some features in dynamic ZIMs.
• JQuery Mode: This mode is now deprecated (and does not in fact require JQuery). It is retained for compatibility with older browsers or frameworks that cannot run Service Workers. The mode has limitations which mean that only static content can be displayed, such as that found in Wikipedia / WikiMedia archives and (for now) Stackexchange. It is a way of inserting articles extracted from the ZIM into the DOM (browser document) by injecting the content into an iframe or tab. We then use native DOM methods, or sometimes JQuery, to parse the article's HTML and insert required assets (images, stylesheets, event listeners for hyperlinks, etc.). On old devices, DOM traversal can be slow, but it is compensated for because we do not extract or run JavaScript assets (which would be technically extremely complicated). As a result, for WikiMedia archives this mode is usually quite fast. On the downside, ZIMs that have a proprietary dynamic UI (such as Gutenberg or TED talks) are only partially supported in this mode: the UI does not work, but articles can be searched for and loaded from the search bar.

↑ Back to Contents

Feedback / helping / contributing

We welcome constructive feedback and suggestions for improvement. We will consider feature requests, but it is even better if you can contribute to improving the app! The source code is on GitHub  , and we welcome Pull Requests (but open an issue and discuss your proposal first).

The bug and issue tracker  is on GitHub too. We use it as our roadmap. Alternatively, you can send your feedback by email. In all cases, if you are reporting a bug, please give as much detail as you can: which device you are using, which version of Kiwix, which archive, which articles etc.

If you like the Kiwix project and use it, please consider giving a donation at https://support.kiwix.org  to help maintain our servers, ZIM files and applications. You can also help by:

• Adding articles to Wikipedia or improving the existing ones, by editing articles ;
• Participating  in the Kiwix project.

↑ Back to Contents

Credits

We owe a big debt of gratitude to:

• Kelson and all the Kiwix community for welcoming us to this project;
• Wikipedia teams and contributors for helping to make knowledge available to everybody;
• All the developers of libraries and tools we used, for releasing them as Free software (see libraries and tools paragraph).

↑ Back to Contents

Licence information

Main libraries and resources used:

• The core-js library, copyright Sebastian McKenzie and contributors, released under an MIT Licence
• The Babel JavaScript compiler, copyright Denis Pushkarev, released under an MIT Licence
• jQuery 3.7, released under the MIT Licence
• Bootstrap 4, released under an MIT Licence
• Fontawesome Free, icons licensed under CC BY 4.0 Licence, code licensed under a MIT Licence
• QUnit 2, released under the MIT Licence
• Kiwix logo from kiwix.org, released under the Creative Commons Attribution Share Alike Licence
• OpenZIM specifications from www.openzim.org, released under the Creative Commons Attribution Share Alike Licence
• Emscripten to generate XZ and ZSTD decompressors in javascript : https://github.com/kripken/emscripten, released under the MIT Licence
• XZ Embedded (the XZ library converted to Javascript with Emscripten), in public domain
• The Zstandard library, released under a BSD Licence and a GPLv2 Copying Licence
• The WebP-Hero browser polyfill, copyright Chase Moskal, released under an MIT Licence
• The http-server, copyright Charlie Robbins, Marak Squires, Jade Michael Thornton and contributors, released under an MIT Licence
• The Rollup JavaScript Bundler, copyright by the contributors, released under an MIT Licence
• Vite Frontend Tooling, copyright Yuxi You and Vite contributors, released under an MIT Licence
• The TestCafé suite, copyright Developer Express Inc., released under an MIT Licence
• Selenium WebDriver, copyright The Software Freedom Conservancy, released under an Apache Lincence v2
• Mocha Testing Framework, copyright The OpenJS Foundation and Mocha contributors, released under a Creative Commons Attribution 4.0 International Licence
• i18next Internationalization Framework, copyright i18next, released under an MIT Licence

↑ Back to Contents

Other platforms/versions

Other Kiwix clients exist on various platforms such as desktop and mobile apps for Linux, Windows, Android and iOS. We also have an offline-first, fully featured PWA based on the code in Kiwix JS: https://pwa.kiwix.org .

See the official download site: https://www.kiwix.org/applications/ .

↑ Back to Contents

Once you have loaded a ZIM archive, you can search its contents using the above search field.