Many of existing criteria for evaluating web sites quality require methods such as heuristic evaluations, or/and empirical usability tests. This paper aims at defining a quality model and a set of characteristics relating internal and external quality factors and giving clues about potential problems, which can be measured by automated tools. The first step in the quality assessment process is an automatic check of the source code, followed by manual evaluation, possibly supported by an appropriate user panel. As many existing tools can check sites (mainly considering accessibility issues), the general architecture will be based upon a conceptual model of the site/page, and the tools will export their output to a Quality Data Base, which is the basis for subsequent actions (checking, reporting test results, etc.).