Conversational agents are computer programmes designed to replicate bidirectional human conversation through spoken or written language, potentially supplemented with nonverbal features. The eCCo database is a searchable repository of primary studies on conversational agents in health and well-being. It catalogs 657 papers currently, published between 1991 and 2022; 51 address the use of medicines. Most of these papers focus on usability rather than rigorous effectiveness or implementation research, underscoring a need for more robust evaluation. The largest category of conversational agents for pharmaceutical use consists of non-embodied agents (n = 24), followed by virtual embodiment only (n = 19), most using virtual humans (n = 16). This database facilitates the comparison and appraisal of existing research in this field, while contributing to a more nuanced understanding of this technology through multidimensional attributes. We aim to enhance the database accuracy and expand its completeness beyond 2022 with the support of the global research community.