WebAnywhere: A Screen Reader On-the-Go
Jeffrey P. Bigham, Craig M. Prince · 2007 · Assets '07: Proceedings of the 9th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/1296843.1296884
Summary
This demonstration paper introduces WebAnywhere, a web-based, self-voicing screen reader that enables blind users to access the web from any computer with a standard browser and sound output, without needing to install specialized software. The paper addresses a fundamental access barrier: while sighted people routinely use computers other than their own — public library terminals, university labs, coffee shop kiosks, airport computers, friends' laptops — blind users are typically locked out of these shared machines because screen readers like JAWS (costing hundreds of dollars) or Window-Eyes are rarely installed on public terminals. Even free alternatives like Fire Vox (a Firefox extension) require installation permissions that users typically lack on shared computers. Portable solutions existed but were expensive: the Freedombox System Access Mobile USB screen reader cost nearly $500, Braille Sense PDAs cost around $5,000, and smartphones with screen reading software cost about $1,000. WebAnywhere solves this by running entirely within a standard web browser with no installation required. Users navigate the DOM of web pages by paragraph, sentence, word, or character, and can quickly jump between headings and other structural elements. The system uses a server-side proxy that fetches requested pages, extracts text, converts it to speech using a text-to-speech engine, and serves audio clips back to the client. Keyboard shortcuts provide navigation control, and prefetching ensures that the next segment of audio is ready before the current one finishes playing.
Key findings
WebAnywhere demonstrated that a fully functional screen reader could be delivered entirely through a web browser without any client-side installation, a novel technical achievement at the time. The system supported core screen reader navigation patterns including DOM traversal at multiple granularities (paragraph, sentence, word, character), heading-based navigation for quick page scanning, and keyboard-driven interaction — all features essential for productive web use by blind people. The architecture used a server-side proxy approach where pages were fetched, text extracted, and speech synthesized on the server, with audio clips streamed back to the browser. A prefetching strategy ensured relatively smooth audio playback by preparing the next speech segment while the current one played. The authors positioned WebAnywhere as serving two audiences: blind end users who needed occasional web access on shared computers, and web developers who wanted a convenient, free tool for testing their sites' accessibility without purchasing commercial screen reader software.
Relevance
WebAnywhere was a pioneering concept in 2007 that challenged the assumption that assistive technology must be installed locally. The idea of a zero-install, web-based screen reader was ahead of its time, anticipating the broader industry shift toward cloud-based and browser-based applications. While the specific technical approach (server-side proxy with synthesized speech) has been largely superseded by built-in screen readers on modern platforms (VoiceOver on macOS/iOS since 2005-2009, TalkBack on Android, Narrator on Windows, and ChromeVox in Chrome OS), the core problem WebAnywhere identified remains relevant: access to assistive technology should not depend on owning a specific device or having installation privileges. For accessibility practitioners, this paper is a reminder that the cost and availability of assistive technology are themselves accessibility barriers. The suggestion that WebAnywhere could also serve as a developer testing tool foreshadowed the later proliferation of accessibility testing tools and the recognition that developers need easy ways to experience their sites as screen reader users do.
Tags: screen readers · web accessibility · blind users · assistive technology · text-to-speech · web-based tools