This study uses a Corpus Assisted Discourse Studies methodology to provide the first systematic analysis of how trust is discursively constructed in crypto-drug markets. The data come from two purpose-built corpora. One comprises all the forum messages posted on the flag ship crypto-drug market Silk Road during the years in which it traded on the hidden net (c. 250 million words). The other corpus comprises all the reports published by the United Nations Office on Drugs and Crime (UNODC) during the same period (c. 153,000 words). Our analysis of trust focuses on the identities of those buying and selling drugs. The findings reveal that the Silk Road community members (a) regularly discussed vendors’ identities alongside a continuum of trust–risk calculation, explicitly identifying both ‘good’ and ‘bad’ practices and hence engaging in self-regulatory discourses, and (b) mainly constructed drug users’ identities in relation to values of expertise, integrity and benevolence. The findings also suggest that hard law enforcement activity, such as crypto-drug market closure, may encourage technological innovation within these markets. Moreover, our results show a disconnect between the discursive reality of the policy-making documents we examined and the very crypto-drug markets that they seek to legislate.