A SpeechWeb is a collection of hyperlinked speech applications, which are accessed remotely by speech browsers running on end-user devices. Links are activated through spoken commands. The first SpeechWeb was developed by Richard A. Frost of the University of Windsor, in Canada, and was demonstrated at the first International Workshop on Advance Issues of E-Commerce and Web-Based Information Systems (Frost 1999a), at the Sixteenth National Conference on Artificial Intelligence (Frost 1999b), and at the 1999 Conference of the Pacific Association for Computational Linguistics PACLING 99 (Frost and Chitte 1999).

A novel architecture for SpeechWebs was developed by the MySpeechWeb research group at the University of Windsor. The new architecture is called the LRRP architecture, standing for Local thin-client application-specific speech Recognition and Remote natural-language query Processing (Frost et al 2004). Users navigate an LRRP SpeechWeb using voice-activated hyperlink commands, and query the knowledge sources through spoken natural-language using a speech browser executing on a local device. When an LRRP SpeechWeb browser is directed to a remote hyperlinked knowledge source it begins by downloading a recognition grammar which is used to tailor the browser to the knowledge contained in the remote application. Queries that are subsequently recognized on the local device are then sent to the remote application for processing. Answers that are returned to the local browser are output as synthesized voice. If a navigation command is spoken, the application returns the address of another application to which the browser is redirected in a manner similar to the following of a link in the conventional web. Grammars can be simple or complex, and applications can be constructed in any programming language.

The notion of a Public-Domain SpeechWeb (PDSW) was first suggested by Richard Frost in a paper published in the Communications of the ACM (Frost 2005). A PDSW is a SpeechWeb which is accessible to the public and which contains speech applications that are created and deployed by the public in a manner that is analogous to the creation and deployment of HTML pages on the conventional web. The MySpeechWeb research group at the University of Windsor, which is funded by the Natural Science and Engineering Research Council of Canada (NSERC), is currently developing software to facilitate the construction and deployment of speech applications by the public.

The most-recent SpeechWeb browser constructed by the MySpeechWeb research group at the University of Windsor, is written in the speech markup language X+V, uses freely-available software, and can be deployed on computers that have the Opera browser in 5 minutes (Frost, Ma, and Shi 2008).