We describe a knowledge server that permits Web users to retrieve and add knowledge in a shared knowledge base. The following features distinguish WebKB-2 from other ontology servers or KBMSs: (i) the ontology is large (at present, 69,000 categories and 87,800 links mostly coming from WordNet) and extendible at any time by any user, (ii) asynchronous cooperation between users is supported and encouraged (users are encouraged to reuse, complement or correct the knowledge of other users but do not have to agree with each other and may add new names to categories) while the knowledge base is kept unique to maximize knowledge interconnection, retrieval and inconsistency detection, (iii) the proposed knowledge representation languages are designed to be both expressive and readable to permit and encourage the users to enter all the knowledge they want (though that still requires motivation). WebKB-2 is ultimately intended to permit cooperatively-built Yellow-Page like catalogs, that is, permit Web users to publish their information in a way that is automatically retrievable and comparable with other users’ knowledge (as opposed to publishing information in plain text documents or even RDF documents). For example, database developpers or car dealers could describe and compare their products in a precise way, supporting precise queries.