Positive species interactions are ubiquitous in natural communities, but the mechanisms through which they operate are poorly understood. One proposed mechanism is resource conversion – the conversion by a benefactor species of a resource from a resource state that is inaccessible to a potential beneficiary species into a resource state that is accessible. Such conversion often occurs as a byproduct of resource consumption, and sometimes in exchange for non‐resource benefits to the benefactor species. At least five known classes of interactions, including both facilitative and mutualistic ones, may be classified as resource conversion interactions. We formulated a generalizable mathematical model for resource conversion interactions and examined two model variants that represent processing chain and nurse plant interactions. We examined the conditions under which these conformed to the stress‐gradient hypothesis (SGH), which predicts increased interaction benefits in more stressful environments. These yielded four key insights: 1) resource conversion interactions can be positive (towards the resource recipient) only when facilitator‐mediated resource conversion is more efficient than the baseline, spontaneous, facilitator‐independent resource conversion; 2) the sign of resource conversion interaction outcomes never switches (e.g. from net positive to net negative) with changing levels of resource availability, when all other parameters are kept constant; 3) processing chain interactions at equilibrium can never be positive in a manner that conforms to the SGH; 4) nurse plant interactions can be positive and conform to the SGH, although the manner in which they do depends largely on how resource stress is defined, and the environmental supply rate of surface soil moisture. The first two insights are likely to be generalizable across all resource conversion interactions. The general agreement of the model with empirical studies suggest that resource conversion is the mechanism underlying the aforementioned interactions, and an ecologically meaningful way of classifying these previously unassociated positive species interactions. [ABSTRACT FROM AUTHOR]