BACKGROUND: Polysubstance use is a defining feature of the opioid overdose epidemic in the United States today. Most quantitative studies of polysubstance use rely on 30-day retrospective measures. Understanding how drugs are co-used in shorter (daily and hourly) timeframes enhances our understanding of polysubstance-related overdose risk. METHODS: We used ecological momentary assessment (EMA) to assess polysubstance in community-based sample of people who use drugs in Oakland, CA (N = 117). Participants provided data on substance use three times daily over 28 days, using a smartphone app. We created binary variables of same-day and same-hour polysubstance use and estimated its prevalence with mixed-effects logistic regression models. RESULTS: The response rate to thrice-daily prompts was 64.6 %. The prevalence of same-day polysubstance use on any given day was 0.77 (95 % CI 0.73, 0.82), and the prevalence of same-hour polysubstance use was 0.71 (95 % CI 0.66, 0.76). Defined as the combination of opioids with stimulants, opioids with alcohol, opioids with benzodiazepines, or more than one opioid in the same hour, the probability of risky polysubstance use on a given day was 0.59 (95 % CI 0.52, 0.65). Excluding methadone from a treatment program, the probability of risky polysubstance use on a given day was 0.51 (95 % CI 0.44, 0.58). CONCLUSIONS: Our findings highlight the importance of targeted prevention efforts addressing the diverse substance combinations that increase overdose risk. It is crucial to develop and test substance use treatment and overdose prevention strategies that address the complexities of polysubstance use, in partnership with people who use drugs.