Indexing the gender of the speaker or the addressee within any type of sentence is often considered as sociolinguistic variation rather than as a gender-exclusive rule. This paper presents a survey of categorical (rather than statistical) gender indexicality in grammar with the greatest number of languages to date. It also offers a data-informed typology of categorical gender indexicality based on 41 indigenous South American languages, aimed at encouraging and facilitating research on genderlects. Examples are classified according to which speech-act participants have their gender indexed and in which area of the grammar (lexicon, discourse markers, phonology, morphology). The main findings are, first, that categorical gender indexicality in the grammar is more frequent than thought and, second, that the weight of gender indexicality within a language is correlated to the domain: it is limited in the lexicon, more present with discourse markers, and pervasive at the phonological and morphological levels. [