Offline-Pedia

4 billion people worldwide lack reliable Internet access -because of costs, lack of infrastructure or outright censorship. In a technology-focused world, this severely impacts an individual’s access to education and better economic opportunities.

Kiwix a universal platform bringing Internet content to people without Internet access. Kiwix copies entire websites and compresses them into single files. These can then be copied and shared at will on computers, mobile phones, or even installed on a low-cost hotspot. Users can browse this content without the need for further Internet connectivity. Though most well-known as the Offline Wikipedia, Kiwix can convert and read any web-based material, in any language. 

Kiwix has been bridging the digital divide in schools in Peru, healthcare centres in the Dominican Republic, and information-seekers in Cuba, and want to provide more people with more, and improved, offline content.

In 2018, 53% of the Latin America and Caribbean region had the infrastructure needed to access mobile Internet data. While this leaves almost half the population without the ability to do so, even those that can often cannot afford the expensive packages offered. For example, The Cuban Telecommunications Enterprise S.A. (ETECSA) started selling 4G mobile data packages in November 2019, but the cost is equivalent to the average monthly salary of state workers. Even when the 3G package was released in 2018, the first time internet data was to be available, even the lowest data package of 600MB cost 7 pesos, a significant chunk of the average monthly wage of 30. 

With such cost and infrastructure problems, there is little room for exploring informational resources. In many rural andean communities, Spanish is also not spoken natively further adding to social and economical imbalances as printed educational resources are limited to the official language. 

Kiwix works to improve the lives of underfunded (and mostly rural) communities who do not have reliable access to Internet resources for digital education purposes. 

Offline-Pedia is based in Ecuador and re-purposes existing hardware (from recycling) and uses Kiwix to create a library of contents that the community would otherwise be unable (or unaware of how) to access.

With this project, Sergio Minniti (a sociologist who works on "critical technical practice" prototypes with students) has been assisting in devising ethnographic studies to further understanding of the needs of those involved with the project. In response to their needs, we plan to provide more content in more in minority languages, including the ability to asynchronously edit Wikipedia.

Kiwix provides an offline library of contents for learning that can be browsed at ease, without data or single-user limitations. Content can be selected from an ever-expanding library of educational websites, meaning that communities can adapt and chose what is offered to them and best suits their needs.

Kiwix can host the entirety of Wikipedia, in any language, as well as sister projects like the Wiktionary, on a single, low-cost microSD card. Other contents of interest include basic word processing software, simple maths and physics games from the University of Colorado, etc.

Kiwix relies on the openZIM technology - a copy of a website is made using this open-source compression format, which allows the creation of an archived version of the desired website as a unique .zim file. It can then by copied, shared and accessed anywhere without further need for an internet connection. The user experience is similar to actually being connected to high-speed internet - except that they can be fully offline.

Offline-pedia is using old CRT screens that would otherwise become electronic waste as the digital switch-over takes place, the entire cost of Offline-Pedia need not cost above $100 per device. 

The Offline-Pedia solution takes a problem of lacking (Internet resources), and uses what it has in abundance (free offline material, unused TV sets) to provide a solution.

The programme then creates an opportunity for all potential users as it can scale pretty seamlessly: each low-cost single-board computer that is used can serve up to 30 people at no additional cost than the initial 100 USD price tag. As it is most often deployed in community centers, the project also offers the social benefit of bringing the community together to share in a resource which ultimately results in large-scale economic advancement; as better learning resources increase, so do employment and education opportunities. 

Moreover, this project offers a self-sustaining, inclusive solution; the community learns how to use and chose the tools it needs to help itself, rather than depending on passively receiving external assistance.

Kiwix has repeatedly seen how offering people a library of offline content they can chose from has in fact catalysed long-term community self empowerment. People are not only treated as consumers of a fossilized corpus of knowledge produced elsewhere, but also prescriptors of what they will decide is best suited for their needs.

Initial deployments of Offline-pedia first tested this theory by showcasing Kiwix and a variety of offline content, making it clear that other options were available (for instance by bringing a second, identical device loaded with other educational resources or languages. This led users to proactively explore further what they wanted from the tool, which ones they liked or not, and select which options were best suited for their learning interests -including the possibility to have content in their own local language (Quechua) rather than just providing the offline readers with off-the-shelf, standardized solutions. In rural areas where what they get is decided by unknown city-based government officials, this is a often a first.

Kiwix and OfflinePedia breaks down barriers not just between the communities it works with and those in Internet-accessing locations, but in fact encourages the sharing of ideas and building of knowledge and education within the community itself.

Kiwix has an estimated user based of 3 to 4 million people in 202 countries and territories in 2019, although due to the nature of offline, this data is difficult to track : we rely on server statistics as well as direct feedback from deployment partners and should be seen as conservative. We have anecdotal evidence from different countries that Kiwis is also distributed laterally (meaning that one person will download the application and contents and share or resell them with friends and neighbours).

Based on current trends and strategy, we expect our user base to roughly double every year meaning 6 million users by the end of this year, and 100 million by 2024-25.

Our short-term goal is to establish long-term partnerships with local actors who can then be leveraged to propagate Kiwix across their own networks. They have a local understanding at their disposal that we can not and never will have, and will best equipped to provide users with a relevant offer, as well as guide our own development efforts by providing structured feedback. Using them as a support to develop and maintain the ubiquitous reach of Kiwix, we will then use it as a tool to engage more owners of online content inclined to offer it offline, showing them than there is more benefit from investing an established, agnostic platform rather than invest in the development of a limited, bespoke solution from scratch. In short, local distribution partners help us increase our user base, which in turns helps us broaden the variety of content offered by attracting new partners (and thus making Kiwix even more enjoyable for more users, etc.).

Kiwix was established in 2017 and are expanding and looking for this to continue; from a 130 kCHF budget in 2017, we grew to 300kCHF in 2018, and look to exceed CHF 1 million by the end of 2020, most of it dedicated towards platform stability and maintenance. With this goal almost reached, the next step for us will be to increase the variety of content on offer at an industrial pace, particularly as far as minority languages are concerned.

Cultural/legal barrier: the limited availability of freely-licensed educational content might be a medium-term barriers as we get to industrialize content harvesting. Things are slowly moving on that front with the growing acceptation of OER and open-access as default release options, but this movement is only starting to take hold in the anglophone world: we do expect some capillarity and for it to make its way into the Spanish, French and Arabic educational resource communities, but this might take quite some time. People still have XXth century view that content restrictions are a great way to maximize audience.

Market barriers: eventhough single-board computers are extremely cheap, trade barriers (custom duties, red tape) in poorer countries are extremely high and block access to equipment, making the choice of distributors more limited that the should be.

There is a global push for open-access, and it is only accelerating, so we are not overly concerned with this. The ultra-dominance of anglophone content and teaching models is more concerning as they do not take into account local specificities. Many a times have we seen local elites publishing their content in English because of the pressure of funding or simply academic / media recognition.

We have developed a cooperative financing model where we try to support would-be distributors with seed funding or resource access.

As mentioned previously, Kiwix is being deployed in Ecuador through the Offline-Pedia project from Yachay Tech.

While this is the main project we are focusing on for this proposal, we are also a resource for a number of other LATAM projects: Columbia University researchers are using Kiwix for schools in the Andes of rural Peru as a hotspot for sharing Internet content, the medical subsets of Wikipedia are being distributed via Kiwix in hospitals across the Dominican Republic, and Kiwix has become household name in Cuba as it is widely available through State-sponsored Joven Club de Computacion (EcuMovil program) and the informal Paquete Semanal.

Connectivity issues remain widespread in the region and are unlikely from being entirely resolved soon, according to the latest GSMA figures in Internet coverage. We plan on continuing to work with our local partners to understand local users' needs, expand our offer in Spanish/Portuguese content and keep on searching for local social entrepreneurs with a vested interest in expanding offline access as a service (bottom up approach).

We also are working with more global partners (e.g. Wikimedia Foundation, Libraries without Borders, etc.) and also plan to expand this more conventional top-down deployment expansion.

9 staff / 6 FTEs  + around 50 volunteers developers from around the world.

As a technology, Kiwix simply is the only content-agnostic solution to bringing internet content to people without connectivity : most other platforms we  are aware of that cater to the offline world are there to sell / promote /distribute the content of their funder alone. This alone helped us become the default solution for a wide range of initiatives, as we have no issue white-branding our product and pass it on as someone else's work: as a non-profit, our mission and vision is to bring knowledge to as many people as possible, a user-oriented philosophy that actually helps us corner the offline market.

On the team side, just under half of Kiwix’ coders are located in the Global South (Mali, Ghana, India) and have hands-on experience with connectivity issues. Kiwix’ advisory committee is also composed of five members representing its main areas of deployment (US, Africa, Middle-East, South-East Asia, South America) and are tasked with providing user and deployment input.

As mentioned previously, we work with many organisations in a non-contractual collaboration capacity, whereby shared information with each other on who Kiwix is serving and for what purpose, and how we can improve the software: we provide them with the software and content to run their operations, and they give us feedback on use cases and audience in return. Some of them may decide to contribute funding specific areas of software development that more closely match their needs. Among these, we have Libraries without Borders (France), the Orange Foundation (France), Fair Denmark (Denmark), eLiS (Germany), Teach-a-Class (India), Offline-Pedia (Ecuador), and many more.

We also have a formal agreement with the Wikimedia Foundation whereby they contribute funds to further enhance offline access to Wikipedia worldwide as is part of their mission (80% of Wikipedia users are in the Global North, whereas 80% of ours are in the South). More specifically, they have assisted in the development of Kiwix desktop and WikiMed (Kiwix's medical subset of Wikipedia).

Our business model relies on three pillars:

- direct sponsorship from like-minded organisations (e.g. Wikimedia);

- IT services for the conversion of content (for a partner interested in deploying specific resources not yet available offline) as well as bespoke software upgrades;

- Sale of pre-loaded, ready-to-deploy hotspot devices via an online hotspot installer.

- In terms of grants, we usually look for multi-year structural grants that allow us to build and maintain our infrastructure. We occasionnally apply for program specific grants if they match our R&D objectives (e.g  a 2019 grant from the Mozilla Foundation to create a universal content scraper);

- Sale of IT services (software development, installing Kiwix on servers) is done on a standard consulting fee basis;

- Sale of equipment and content is done through a variety of local distributors so as to save on shipping costs and import duties, meaning that we provide them with content that they are allowed to sell at the price they choose, with a 10% cut being sent back to us.

Our reasons for applying are simple: we believe that the TPrize could help us further our cause greatly, and we believe that our proposal fits.

We meet the TPrize's criteria in that we have a successful solution to our problem (lack of Internet connectivity affecting livelihoods globally) which has been successfully implemented, and has grown significantly since its implementation. We are now looking to scale our solution so that it can grow in range (of educational material we provide) and spread (who we provide it to). Our proposal is also simple, yet highly effective, and easily replicable for a variety of different contexts. 

The offer of connection with experts would be a huge benefit to the deployment and development of Kiwix. As a non-profit, our ultimate aim is to connect the unconnected and to ensure that our software satisfies the needs of those who are most lacking in Internet resources. To have expert advice (be it technical, administrative, communicative, or other) would help us to streamline our processes and get our software where it needs to be more effectively. 

We want to work with educational and/or digital solutions organisations, who could use Kiwix as a tool to bring Internet content as an informational tool for advancement of a community.